├── .gitignore
├── archive
    ├── readme.md
    └── word-movers-distance-exploration.ipynb
├── bicluster.xlsx
├── biclustering-with-paul-revere.ipynb
├── binder
    ├── postBuild
    └── requirements.txt
├── chart.html
├── cleaned_hm.csv.xz
├── colored-roc-curves.ipynb
├── dimension-reduction-umap-to-plotly-scatterplot.ipynb
├── docs
    ├── nlp-sbert-umap-hdbscan-tsss-sphere-2d.html
    ├── nlp-sbert-umap-tsss-hdbscan-chart.html
    ├── nlp-sbert-umap-tsss-hdbscan-sphere.html
    └── nlp-use-umap-hdbscan-chart.html
├── holy-nlp.ipynb
├── nlp-sbert-umap-hdbscan.ipynb
├── nlp-use-umap-hdbscan.ipynb
├── output-transformations-roc-impact.ipynb
├── pipfreeze.sh
├── readme.md
├── roc_colored.png
├── rule-based-matching-with-spacy-matcher.ipynb
└── w2v_tsne.csv


/.gitignore:
--------------------------------------------------------------------------------
  1 | # Byte-compiled / optimized / DLL files
  2 | __pycache__/
  3 | *.py[cod]
  4 | *$py.class
  5 | 
  6 | # C extensions
  7 | *.so
  8 | 
  9 | # Distribution / packaging
 10 | .Python
 11 | build/
 12 | develop-eggs/
 13 | dist/
 14 | downloads/
 15 | eggs/
 16 | .eggs/
 17 | lib/
 18 | lib64/
 19 | parts/
 20 | sdist/
 21 | var/
 22 | wheels/
 23 | *.egg-info/
 24 | .installed.cfg
 25 | *.egg
 26 | MANIFEST
 27 | 
 28 | # PyInstaller
 29 | #  Usually these files are written by a python script from a template
 30 | #  before PyInstaller builds the exe, so as to inject date/other infos into it.
 31 | *.manifest
 32 | *.spec
 33 | 
 34 | # Installer logs
 35 | pip-log.txt
 36 | pip-delete-this-directory.txt
 37 | 
 38 | # Unit test / coverage reports
 39 | htmlcov/
 40 | .tox/
 41 | .coverage
 42 | .coverage.*
 43 | .cache
 44 | nosetests.xml
 45 | coverage.xml
 46 | *.cover
 47 | .hypothesis/
 48 | .pytest_cache/
 49 | 
 50 | # Translations
 51 | *.mo
 52 | *.pot
 53 | 
 54 | # Django stuff:
 55 | *.log
 56 | local_settings.py
 57 | db.sqlite3
 58 | 
 59 | # Flask stuff:
 60 | instance/
 61 | .webassets-cache
 62 | 
 63 | # Scrapy stuff:
 64 | .scrapy
 65 | 
 66 | # Sphinx documentation
 67 | docs/_build/
 68 | 
 69 | # PyBuilder
 70 | target/
 71 | 
 72 | # Jupyter Notebook
 73 | .ipynb_checkpoints
 74 | 
 75 | # pyenv
 76 | .python-version
 77 | 
 78 | # celery beat schedule file
 79 | celerybeat-schedule
 80 | 
 81 | # SageMath parsed files
 82 | *.sage.py
 83 | 
 84 | # Environments
 85 | .env
 86 | .venv
 87 | env/
 88 | venv/
 89 | ENV/
 90 | env.bak/
 91 | venv.bak/
 92 | 
 93 | # Spyder project settings
 94 | .spyderproject
 95 | .spyproject
 96 | 
 97 | # Rope project settings
 98 | .ropeproject
 99 | 
100 | # mkdocs documentation
101 | /site
102 | 
103 | # mypy
104 | .mypy_cache/
105 | 
106 | *.gz
107 | .vscode/
108 | .DS_Store


--------------------------------------------------------------------------------
/archive/readme.md:
--------------------------------------------------------------------------------
1 | # The Archive
2 | 
3 | Notebooks here are not guaranteed to work due to outdated dependencies or other issues.


--------------------------------------------------------------------------------
/bicluster.xlsx:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/pmbaumgartner/binder-notebooks/cfee899187c9ba5e7a11a5b26fb37aa78cbc1201/bicluster.xlsx


--------------------------------------------------------------------------------
/biclustering-with-paul-revere.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "markdown",
  5 |    "metadata": {},
  6 |    "source": [
  7 |     "# Biclustering (with Paul Revere)\n",
  8 |     "\n",
  9 |     "\n",
 10 |     "ref: http://scikit-learn.org/stable/auto_examples/bicluster/bicluster_newsgroups.html"
 11 |    ]
 12 |   },
 13 |   {
 14 |    "cell_type": "markdown",
 15 |    "metadata": {},
 16 |    "source": [
 17 |     "Biclustering is a helpful clustering procedure that derives cluster membership for both rows and columns. Clustering rows and columns individually is fairly common. I view biclustering as an extension that allows me to ask: \"If these columns are clustered together, which rows (if any) point towards that clustering? Accordingly, if any rows are clustered together, which column values might tell me why they're clustered?\"\n",
 18 |     "\n",
 19 |     "I have found it helpful with:\n",
 20 |     "- Incidence Matrix / Cooccurence data (typically found before converting something into a graph/network structure)\n",
 21 |     "- Text data (the [original paper](http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.140.3011&rep=rep1&type=pdf) on coclustering does this)\n",
 22 |     "\n",
 23 |     "The example below performs coclustering on the awesome \"Paul Revere dataset\" originally analyzed in an [amazing blog post](https://kieranhealy.org/blog/archives/2013/06/09/using-metadata-to-find-paul-revere/) by [Kieran Healy](https://kieranhealy.org/).\n",
 24 |     "\n",
 25 |     "## References:\n",
 26 |     "> Biclustering algorithms simultaneously cluster rows and columns of a data matrix. These clusters of rows and columns are known as biclusters. Each determines a submatrix of the original data matrix with some desired properties.\n",
 27 |     "\n",
 28 |     "[sklearn - Biclustering](http://scikit-learn.org/stable/modules/biclustering.html#biclustering)"
 29 |    ]
 30 |   },
 31 |   {
 32 |    "cell_type": "code",
 33 |    "execution_count": 1,
 34 |    "metadata": {},
 35 |    "outputs": [],
 36 |    "source": [
 37 |     "import re\n",
 38 |     "\n",
 39 |     "from sklearn.datasets.twenty_newsgroups import fetch_20newsgroups\n",
 40 |     "from sklearn.feature_extraction.text import TfidfVectorizer\n",
 41 |     "from sklearn.cluster.bicluster import SpectralCoclustering\n",
 42 |     "\n",
 43 |     "import pandas as pd\n",
 44 |     "import numpy as np\n",
 45 |     "import matplotlib.pylab as plt\n",
 46 |     "\n",
 47 |     "%matplotlib inline"
 48 |    ]
 49 |   },
 50 |   {
 51 |    "cell_type": "markdown",
 52 |    "metadata": {},
 53 |    "source": [
 54 |     "## Data Import"
 55 |    ]
 56 |   },
 57 |   {
 58 |    "cell_type": "code",
 59 |    "execution_count": 2,
 60 |    "metadata": {},
 61 |    "outputs": [],
 62 |    "source": [
 63 |     "data = pd.read_csv(\n",
 64 |     "    \"https://raw.githubusercontent.com/kjhealy/revere/master/data/PaulRevereAppD.csv\",\n",
 65 |     "    index_col=0,\n",
 66 |     ")"
 67 |    ]
 68 |   },
 69 |   {
 70 |    "cell_type": "code",
 71 |    "execution_count": 3,
 72 |    "metadata": {},
 73 |    "outputs": [
 74 |     {
 75 |      "data": {
 76 |       "text/html": [
 77 |        "<div>\n",
 78 |        "<style scoped>\n",
 79 |        "    .dataframe tbody tr th:only-of-type {\n",
 80 |        "        vertical-align: middle;\n",
 81 |        "    }\n",
 82 |        "\n",
 83 |        "    .dataframe tbody tr th {\n",
 84 |        "        vertical-align: top;\n",
 85 |        "    }\n",
 86 |        "\n",
 87 |        "    .dataframe thead th {\n",
 88 |        "        text-align: right;\n",
 89 |        "    }\n",
 90 |        "</style>\n",
 91 |        "<table border=\"1\" class=\"dataframe\">\n",
 92 |        "  <thead>\n",
 93 |        "    <tr style=\"text-align: right;\">\n",
 94 |        "      <th></th>\n",
 95 |        "      <th>StAndrewsLodge</th>\n",
 96 |        "      <th>LoyalNine</th>\n",
 97 |        "      <th>NorthCaucus</th>\n",
 98 |        "      <th>LongRoomClub</th>\n",
 99 |        "      <th>TeaParty</th>\n",
100 |        "      <th>BostonCommittee</th>\n",
101 |        "      <th>LondonEnemies</th>\n",
102 |        "    </tr>\n",
103 |        "  </thead>\n",
104 |        "  <tbody>\n",
105 |        "    <tr>\n",
106 |        "      <th>Adams.John</th>\n",
107 |        "      <td>0</td>\n",
108 |        "      <td>0</td>\n",
109 |        "      <td>1</td>\n",
110 |        "      <td>1</td>\n",
111 |        "      <td>0</td>\n",
112 |        "      <td>0</td>\n",
113 |        "      <td>0</td>\n",
114 |        "    </tr>\n",
115 |        "    <tr>\n",
116 |        "      <th>Adams.Samuel</th>\n",
117 |        "      <td>0</td>\n",
118 |        "      <td>0</td>\n",
119 |        "      <td>1</td>\n",
120 |        "      <td>1</td>\n",
121 |        "      <td>0</td>\n",
122 |        "      <td>1</td>\n",
123 |        "      <td>1</td>\n",
124 |        "    </tr>\n",
125 |        "    <tr>\n",
126 |        "      <th>Allen.Dr</th>\n",
127 |        "      <td>0</td>\n",
128 |        "      <td>0</td>\n",
129 |        "      <td>1</td>\n",
130 |        "      <td>0</td>\n",
131 |        "      <td>0</td>\n",
132 |        "      <td>0</td>\n",
133 |        "      <td>0</td>\n",
134 |        "    </tr>\n",
135 |        "    <tr>\n",
136 |        "      <th>Appleton.Nathaniel</th>\n",
137 |        "      <td>0</td>\n",
138 |        "      <td>0</td>\n",
139 |        "      <td>1</td>\n",
140 |        "      <td>0</td>\n",
141 |        "      <td>0</td>\n",
142 |        "      <td>1</td>\n",
143 |        "      <td>0</td>\n",
144 |        "    </tr>\n",
145 |        "    <tr>\n",
146 |        "      <th>Ash.Gilbert</th>\n",
147 |        "      <td>1</td>\n",
148 |        "      <td>0</td>\n",
149 |        "      <td>0</td>\n",
150 |        "      <td>0</td>\n",
151 |        "      <td>0</td>\n",
152 |        "      <td>0</td>\n",
153 |        "      <td>0</td>\n",
154 |        "    </tr>\n",
155 |        "  </tbody>\n",
156 |        "</table>\n",
157 |        "</div>"
158 |       ],
159 |       "text/plain": [
160 |        "                    StAndrewsLodge  LoyalNine  NorthCaucus  LongRoomClub  \\\n",
161 |        "Adams.John                       0          0            1             1   \n",
162 |        "Adams.Samuel                     0          0            1             1   \n",
163 |        "Allen.Dr                         0          0            1             0   \n",
164 |        "Appleton.Nathaniel               0          0            1             0   \n",
165 |        "Ash.Gilbert                      1          0            0             0   \n",
166 |        "\n",
167 |        "                    TeaParty  BostonCommittee  LondonEnemies  \n",
168 |        "Adams.John                 0                0              0  \n",
169 |        "Adams.Samuel               0                1              1  \n",
170 |        "Allen.Dr                   0                0              0  \n",
171 |        "Appleton.Nathaniel         0                1              0  \n",
172 |        "Ash.Gilbert                0                0              0  "
173 |       ]
174 |      },
175 |      "execution_count": 3,
176 |      "metadata": {},
177 |      "output_type": "execute_result"
178 |     }
179 |    ],
180 |    "source": [
181 |     "data.head()"
182 |    ]
183 |   },
184 |   {
185 |    "cell_type": "code",
186 |    "execution_count": 4,
187 |    "metadata": {},
188 |    "outputs": [],
189 |    "source": [
190 |     "data.index = [i.replace(\".\", \", \") for i in data.index]"
191 |    ]
192 |   },
193 |   {
194 |    "cell_type": "code",
195 |    "execution_count": 5,
196 |    "metadata": {},
197 |    "outputs": [
198 |     {
199 |      "name": "stdout",
200 |      "output_type": "stream",
201 |      "text": [
202 |       "rows: 254 columns: 7\n"
203 |      ]
204 |     }
205 |    ],
206 |    "source": [
207 |     "nrows, ncolumns = data.shape\n",
208 |     "print(\"rows:\", nrows, \"columns:\", ncolumns)"
209 |    ]
210 |   },
211 |   {
212 |    "cell_type": "markdown",
213 |    "metadata": {},
214 |    "source": [
215 |     "## Fit Model"
216 |    ]
217 |   },
218 |   {
219 |    "cell_type": "code",
220 |    "execution_count": 6,
221 |    "metadata": {},
222 |    "outputs": [
223 |     {
224 |      "data": {
225 |       "text/plain": [
226 |        "SpectralCoclustering(init='k-means++', mini_batch=False, n_clusters=6,\n",
227 |        "                     n_init=10, n_jobs=None, n_svd_vecs=None, random_state=666,\n",
228 |        "                     svd_method='arpack')"
229 |       ]
230 |      },
231 |      "execution_count": 6,
232 |      "metadata": {},
233 |      "output_type": "execute_result"
234 |     }
235 |    ],
236 |    "source": [
237 |     "cocluster = SpectralCoclustering(n_clusters=6, random_state=666, svd_method=\"arpack\")\n",
238 |     "cocluster.fit(data)"
239 |    ]
240 |   },
241 |   {
242 |    "cell_type": "markdown",
243 |    "metadata": {},
244 |    "source": [
245 |     "## Explore Results"
246 |    ]
247 |   },
248 |   {
249 |    "cell_type": "code",
250 |    "execution_count": 7,
251 |    "metadata": {},
252 |    "outputs": [
253 |     {
254 |      "name": "stdout",
255 |      "output_type": "stream",
256 |      "text": [
257 |       "Cluster:  0 \tMembers: 68 \tOrgs: 3\n",
258 |       "Orgs:    LongRoomClub; BostonCommittee; LondonEnemies\n",
259 |       "Members: Adams, John; Adams, Samuel; Appleton, Nathaniel; Austin, Benjamin; Austin, Samuel; Baldwin, Cyrus; Boyer, Peter; Boynton, Richard; Brackett, Jos; Bradford, John; Brimmer, Herman; Brimmer, Martin; Broomfield, Henry; Brown, Enoch; Brown, John; Cheever, Ezekiel; Church, Benjamin; Cooper, William; Davis, Caleb; Davis, Edward; Davis, William; Dawes, Thomas; Dennie, William; Dexter, Samuel; Fleet, Thomas; Foster, Bos; Gill, Moses; Greenleaf, Joseph; Greenleaf, William; Greenough, Newn; Hancock, Eben; Hancock, John; Hill, Alexander; Hopkins, Caleb; Isaac, Pierce; Ivers, James; Jarvis, Charles; Johnston, Eben; Lambert, John; Mackay, William; Marshall, Thomas; Marson, John; Mason, Jonathan; Noyces, Nat; Otis, James; Parkman, Elias; Partridge, Sam; Phillips, Samuel; Phillips, William; Pierpont, Robert; Pitts, John; Pitts, Samuel; Powell, William; Prince, Job; Pulling, John; Quincy, Josiah; Roylson, Thomas; Ruddock, Abiel; Sweetser, John; Tyler, Royall; Vernon, Fortesque; Waldo, Benjamin; Wendell, Oliver; Whitwell, Samuel; Whitwell, William; Williams, Jonathan; Winslow, John; Winthrop, John\n",
260 |       "---\n",
261 |       "Cluster:  1 \tMembers: 81 \tOrgs: 1\n",
262 |       "Orgs:    TeaParty\n",
263 |       "Members: Barnard, Samuel; Bolter, Thomas; Bradlee, David; Bradlee, Josiah; Bradlee, Nathaniel; Bradlee, Thomas; Bewer, James; Bruce, Stephen; Burton, Benjamin; Campbell, Nicholas; Clarke, Benjamin; Cochran, John; Colesworthy, Gilbert; Collier, Gershom; Crane, John; Davis, Robert; Dolbear, Edward; Eaton, Joseph; Eckley, Unknown; Etheridge, William; Fenno, Samuel; Foster, Samuel; Frothingham, Nathaniel; Gammell, John; Gore, Samuel; Greene, Nathaniel; Hammond, Samuel; Hendley, William; Hewes, George; Hicks, John; Hobbs, Samuel; Hooton, John; Howard, Samuel; Howe, Edward; Hunnewell, Jonathan; Hunnewell, Richard; Hunstable, Thomas; Hunt, Abraham; Ingersoll, Daniel; Kinnison, David; Lee, Joseph; Lincoln, Amos; Loring, Matthew; Machin, Thomas; MacKintosh, Capt; MacNeil, Archibald; May, John; Melville, Thomas; Moore, Thomas; Morse, Anthony; Mountford, Joseph; Newell, Eliphelet; Palmer, Joseph; Parker, Jonathan; Payson, Joseph; Peters, John; Pierce, William; Pitts, Lendall; Porter, Thomas; Prentiss, Henry; Prince, John; Purkitt, Henry; Randall, John; Roby, Joseph; Russell, John; Russell, William; Sessions, Robert; Shed, Joseph; Simpson, Benjamin; Slater, Peter; Spear, Thomas; Sprague, Samuel; Spurr, John; Starr, James; Stearns, Phineas; Stevens, Ebenezer; Wheeler, Josiah; Williams, Jeremiah; Williams, Thomas; Willis, Nathaniel; Wyeth, Joshua\n",
264 |       "---\n",
265 |       "Cluster:  2 \tMembers: 46 \tOrgs: 1\n",
266 |       "Orgs:    StAndrewsLodge\n",
267 |       "Members: Ash, Gilbert; Bell, William; Blake, Increase; Bray, George; Brown, Hugh; Burbeck, Edward; Burbeck, William; Cailleteau, Edward; Callendar, Elisha; Chipman, Seth; Collins, Ezra; Deshon, Moses; Doyle, Peter; Ferrell, Ambrose; Flagg, Josiah; Gould, William; Graham, James; Gray, Wait; Ham, William; Hitchborn, Nathaniel; Hoffins, John; Inglish, Alexander; Jarvis, Edward; Jefferds, Unknown; Jenkins, John; Kerr, Walter; Lewis, Phillip; Marett, Phillip; Marlton, John; McAlpine, William; Milliken, Thomas; Moody, Samuel; Nicholls, Unknown; Obear, Israel; Palfrey, William; Phillips, John; Potter, Edward; Pulling, Richard; Seward, James; Sloper, Ambrose; Stanbridge, Henry; Tabor, Philip; Webb, Joseph; Webster, Thomas; Whitten, John; Wingfield, William\n",
268 |       "---\n",
269 |       "Cluster:  3 \tMembers: 12 \tOrgs: 0\n",
270 |       "Orgs:    \n",
271 |       "Members: Bass, Henry; Chase, Thomas; Collson, Adam; Condy, JamesFoster; Cooper, Samuel; Eayres, Joseph; Grant, Moses; Molineux, William; Proctor, Edward; Story, Elisha; Swan, James; Young, Thomas\n",
272 |       "---\n",
273 |       "Cluster:  4 \tMembers: 37 \tOrgs: 1\n",
274 |       "Orgs:    NorthCaucus\n",
275 |       "Members: Allen, Dr; Avery, John; Ballard, John; Barber, Nathaniel; Boit, John; Breck, William; Burt, Benjamin; Cazneau, Capt; Chadwell, Mr; Champney, Caleb; Chrysty, Thomas; Edes, Benjamin; Emmes, Samuel; Hickling, William; Hitchborn, Thomas; Holmes, Nathaniel; Hoskins, William; Johonnott, Gabriel; Kent, Benjamin; Kimball, Thomas; Lowell, John; Matchett, John; Merrit, John; Morton, Perez; Palms, Richard; Pearce, IsaacJun; Pearce, Isaac; Peck, Thomas; Sharp, Gibbens; Sigourney, John; Stoddard, Asa; Stoddard, Jonathan; Symmes, Eben; Symmes, John; Tileston, Thomas; Warren, Joseph; White, Samuel\n",
276 |       "---\n",
277 |       "Cluster:  5 \tMembers: 10 \tOrgs: 1\n",
278 |       "Orgs:    LoyalNine\n",
279 |       "Members: Barrett, Samuel; Cleverly, Stephen; Crafts, Thomas; Field, Joseph; Peck, Samuel; Revere, Paul; Smith, John; Trott, George; Urann, Thomas; Welles, Henry\n",
280 |       "---\n"
281 |      ]
282 |     }
283 |    ],
284 |    "source": [
285 |     "cluster_data = []\n",
286 |     "for cluster in range(cocluster.n_clusters):\n",
287 |     "    row_ix, col_ix = cocluster.get_indices(cluster)\n",
288 |     "    data_subset = data.iloc[row_ix, col_ix]\n",
289 |     "    c_rows, c_cols = data_subset.shape\n",
290 |     "    orgs = data_subset.columns\n",
291 |     "    names = data_subset.index\n",
292 |     "    print(\"Cluster: \", cluster, \"\\tMembers:\", c_rows, \"\\tOrgs:\", c_cols)\n",
293 |     "    print(\"Orgs:   \", \"; \".join([\"{}\".format(org) for org in orgs]))\n",
294 |     "    print(\"Members:\", \"; \".join([\"{}\".format(name) for name in names]))\n",
295 |     "    print(\"---\")\n",
296 |     "    cluster = {\n",
297 |     "        \"Cluster\": cluster,\n",
298 |     "        \"Members\": \"; \".join([\"{}\".format(name) for name in names]),\n",
299 |     "        \"Orgs\": \"; \".join([\"{}\".format(org) for org in orgs]),\n",
300 |     "    }\n",
301 |     "    cluster_data.append(cluster)"
302 |    ]
303 |   },
304 |   {
305 |    "cell_type": "code",
306 |    "execution_count": 8,
307 |    "metadata": {},
308 |    "outputs": [],
309 |    "source": [
310 |     "cluster_info = pd.DataFrame(cluster_data)"
311 |    ]
312 |   },
313 |   {
314 |    "cell_type": "code",
315 |    "execution_count": 9,
316 |    "metadata": {},
317 |    "outputs": [
318 |     {
319 |      "data": {
320 |       "text/html": [
321 |        "<div>\n",
322 |        "<style scoped>\n",
323 |        "    .dataframe tbody tr th:only-of-type {\n",
324 |        "        vertical-align: middle;\n",
325 |        "    }\n",
326 |        "\n",
327 |        "    .dataframe tbody tr th {\n",
328 |        "        vertical-align: top;\n",
329 |        "    }\n",
330 |        "\n",
331 |        "    .dataframe thead th {\n",
332 |        "        text-align: right;\n",
333 |        "    }\n",
334 |        "</style>\n",
335 |        "<table border=\"1\" class=\"dataframe\">\n",
336 |        "  <thead>\n",
337 |        "    <tr style=\"text-align: right;\">\n",
338 |        "      <th></th>\n",
339 |        "      <th>Cluster</th>\n",
340 |        "      <th>Members</th>\n",
341 |        "      <th>Orgs</th>\n",
342 |        "    </tr>\n",
343 |        "  </thead>\n",
344 |        "  <tbody>\n",
345 |        "    <tr>\n",
346 |        "      <th>0</th>\n",
347 |        "      <td>0</td>\n",
348 |        "      <td>Adams, John; Adams, Samuel; Appleton, Nathanie...</td>\n",
349 |        "      <td>LongRoomClub; BostonCommittee; LondonEnemies</td>\n",
350 |        "    </tr>\n",
351 |        "    <tr>\n",
352 |        "      <th>1</th>\n",
353 |        "      <td>1</td>\n",
354 |        "      <td>Barnard, Samuel; Bolter, Thomas; Bradlee, Davi...</td>\n",
355 |        "      <td>TeaParty</td>\n",
356 |        "    </tr>\n",
357 |        "    <tr>\n",
358 |        "      <th>2</th>\n",
359 |        "      <td>2</td>\n",
360 |        "      <td>Ash, Gilbert; Bell, William; Blake, Increase; ...</td>\n",
361 |        "      <td>StAndrewsLodge</td>\n",
362 |        "    </tr>\n",
363 |        "    <tr>\n",
364 |        "      <th>3</th>\n",
365 |        "      <td>3</td>\n",
366 |        "      <td>Bass, Henry; Chase, Thomas; Collson, Adam; Con...</td>\n",
367 |        "      <td></td>\n",
368 |        "    </tr>\n",
369 |        "    <tr>\n",
370 |        "      <th>4</th>\n",
371 |        "      <td>4</td>\n",
372 |        "      <td>Allen, Dr; Avery, John; Ballard, John; Barber,...</td>\n",
373 |        "      <td>NorthCaucus</td>\n",
374 |        "    </tr>\n",
375 |        "    <tr>\n",
376 |        "      <th>5</th>\n",
377 |        "      <td>5</td>\n",
378 |        "      <td>Barrett, Samuel; Cleverly, Stephen; Crafts, Th...</td>\n",
379 |        "      <td>LoyalNine</td>\n",
380 |        "    </tr>\n",
381 |        "  </tbody>\n",
382 |        "</table>\n",
383 |        "</div>"
384 |       ],
385 |       "text/plain": [
386 |        "   Cluster                                            Members  \\\n",
387 |        "0        0  Adams, John; Adams, Samuel; Appleton, Nathanie...   \n",
388 |        "1        1  Barnard, Samuel; Bolter, Thomas; Bradlee, Davi...   \n",
389 |        "2        2  Ash, Gilbert; Bell, William; Blake, Increase; ...   \n",
390 |        "3        3  Bass, Henry; Chase, Thomas; Collson, Adam; Con...   \n",
391 |        "4        4  Allen, Dr; Avery, John; Ballard, John; Barber,...   \n",
392 |        "5        5  Barrett, Samuel; Cleverly, Stephen; Crafts, Th...   \n",
393 |        "\n",
394 |        "                                           Orgs  \n",
395 |        "0  LongRoomClub; BostonCommittee; LondonEnemies  \n",
396 |        "1                                      TeaParty  \n",
397 |        "2                                StAndrewsLodge  \n",
398 |        "3                                                \n",
399 |        "4                                   NorthCaucus  \n",
400 |        "5                                     LoyalNine  "
401 |       ]
402 |      },
403 |      "execution_count": 9,
404 |      "metadata": {},
405 |      "output_type": "execute_result"
406 |     }
407 |    ],
408 |    "source": [
409 |     "cluster_info"
410 |    ]
411 |   },
412 |   {
413 |    "cell_type": "markdown",
414 |    "metadata": {},
415 |    "source": [
416 |     "## Format and Sort DataFrame with MultiIndex\n",
417 |     "Since rows and columns now have hierarchical information with cluster membership, we can reindex the dataframe to take this into account and index easily on cluster. This should allow us to see the block cluster structure along the diagonal."
418 |    ]
419 |   },
420 |   {
421 |    "cell_type": "code",
422 |    "execution_count": 10,
423 |    "metadata": {},
424 |    "outputs": [],
425 |    "source": [
426 |     "data_mi = data.copy()\n",
427 |     "\n",
428 |     "cocluster_rows = list(zip(cocluster.row_labels_, data.index))\n",
429 |     "row_index = pd.MultiIndex.from_tuples(cocluster_rows, names=[\"cluster\", \"person\"])\n",
430 |     "data_mi.index = row_index\n",
431 |     "data_mi = data_mi.sort_index(axis=0)\n",
432 |     "\n",
433 |     "cocluster_columns = list(zip(cocluster.column_labels_, data.columns))\n",
434 |     "col_index = pd.MultiIndex.from_tuples(cocluster_columns, names=[\"cluster\", \"org\"])\n",
435 |     "data_mi.columns = col_index\n",
436 |     "data_mi = data_mi.sort_index(axis=1)"
437 |    ]
438 |   },
439 |   {
440 |    "cell_type": "code",
441 |    "execution_count": 11,
442 |    "metadata": {},
443 |    "outputs": [
444 |     {
445 |      "data": {
446 |       "text/html": [
447 |        "<div>\n",
448 |        "<style scoped>\n",
449 |        "    .dataframe tbody tr th:only-of-type {\n",
450 |        "        vertical-align: middle;\n",
451 |        "    }\n",
452 |        "\n",
453 |        "    .dataframe tbody tr th {\n",
454 |        "        vertical-align: top;\n",
455 |        "    }\n",
456 |        "\n",
457 |        "    .dataframe thead tr th {\n",
458 |        "        text-align: left;\n",
459 |        "    }\n",
460 |        "\n",
461 |        "    .dataframe thead tr:last-of-type th {\n",
462 |        "        text-align: right;\n",
463 |        "    }\n",
464 |        "</style>\n",
465 |        "<table border=\"1\" class=\"dataframe\">\n",
466 |        "  <thead>\n",
467 |        "    <tr>\n",
468 |        "      <th></th>\n",
469 |        "      <th>cluster</th>\n",
470 |        "      <th colspan=\"3\" halign=\"left\">0</th>\n",
471 |        "      <th>1</th>\n",
472 |        "      <th>2</th>\n",
473 |        "      <th>4</th>\n",
474 |        "      <th>5</th>\n",
475 |        "    </tr>\n",
476 |        "    <tr>\n",
477 |        "      <th></th>\n",
478 |        "      <th>org</th>\n",
479 |        "      <th>BostonCommittee</th>\n",
480 |        "      <th>LondonEnemies</th>\n",
481 |        "      <th>LongRoomClub</th>\n",
482 |        "      <th>TeaParty</th>\n",
483 |        "      <th>StAndrewsLodge</th>\n",
484 |        "      <th>NorthCaucus</th>\n",
485 |        "      <th>LoyalNine</th>\n",
486 |        "    </tr>\n",
487 |        "    <tr>\n",
488 |        "      <th>cluster</th>\n",
489 |        "      <th>person</th>\n",
490 |        "      <th></th>\n",
491 |        "      <th></th>\n",
492 |        "      <th></th>\n",
493 |        "      <th></th>\n",
494 |        "      <th></th>\n",
495 |        "      <th></th>\n",
496 |        "      <th></th>\n",
497 |        "    </tr>\n",
498 |        "  </thead>\n",
499 |        "  <tbody>\n",
500 |        "    <tr>\n",
501 |        "      <th rowspan=\"10\" valign=\"top\">0</th>\n",
502 |        "      <th>Adams, John</th>\n",
503 |        "      <td>0</td>\n",
504 |        "      <td>0</td>\n",
505 |        "      <td>1</td>\n",
506 |        "      <td>0</td>\n",
507 |        "      <td>0</td>\n",
508 |        "      <td>1</td>\n",
509 |        "      <td>0</td>\n",
510 |        "    </tr>\n",
511 |        "    <tr>\n",
512 |        "      <th>Adams, Samuel</th>\n",
513 |        "      <td>1</td>\n",
514 |        "      <td>1</td>\n",
515 |        "      <td>1</td>\n",
516 |        "      <td>0</td>\n",
517 |        "      <td>0</td>\n",
518 |        "      <td>1</td>\n",
519 |        "      <td>0</td>\n",
520 |        "    </tr>\n",
521 |        "    <tr>\n",
522 |        "      <th>Appleton, Nathaniel</th>\n",
523 |        "      <td>1</td>\n",
524 |        "      <td>0</td>\n",
525 |        "      <td>0</td>\n",
526 |        "      <td>0</td>\n",
527 |        "      <td>0</td>\n",
528 |        "      <td>1</td>\n",
529 |        "      <td>0</td>\n",
530 |        "    </tr>\n",
531 |        "    <tr>\n",
532 |        "      <th>Austin, Benjamin</th>\n",
533 |        "      <td>0</td>\n",
534 |        "      <td>1</td>\n",
535 |        "      <td>0</td>\n",
536 |        "      <td>0</td>\n",
537 |        "      <td>0</td>\n",
538 |        "      <td>0</td>\n",
539 |        "      <td>0</td>\n",
540 |        "    </tr>\n",
541 |        "    <tr>\n",
542 |        "      <th>Austin, Samuel</th>\n",
543 |        "      <td>0</td>\n",
544 |        "      <td>1</td>\n",
545 |        "      <td>0</td>\n",
546 |        "      <td>0</td>\n",
547 |        "      <td>0</td>\n",
548 |        "      <td>0</td>\n",
549 |        "      <td>0</td>\n",
550 |        "    </tr>\n",
551 |        "    <tr>\n",
552 |        "      <th>Baldwin, Cyrus</th>\n",
553 |        "      <td>0</td>\n",
554 |        "      <td>1</td>\n",
555 |        "      <td>0</td>\n",
556 |        "      <td>0</td>\n",
557 |        "      <td>0</td>\n",
558 |        "      <td>0</td>\n",
559 |        "      <td>0</td>\n",
560 |        "    </tr>\n",
561 |        "    <tr>\n",
562 |        "      <th>Boyer, Peter</th>\n",
563 |        "      <td>0</td>\n",
564 |        "      <td>1</td>\n",
565 |        "      <td>0</td>\n",
566 |        "      <td>0</td>\n",
567 |        "      <td>0</td>\n",
568 |        "      <td>0</td>\n",
569 |        "      <td>0</td>\n",
570 |        "    </tr>\n",
571 |        "    <tr>\n",
572 |        "      <th>Boynton, Richard</th>\n",
573 |        "      <td>1</td>\n",
574 |        "      <td>1</td>\n",
575 |        "      <td>0</td>\n",
576 |        "      <td>0</td>\n",
577 |        "      <td>0</td>\n",
578 |        "      <td>0</td>\n",
579 |        "      <td>0</td>\n",
580 |        "    </tr>\n",
581 |        "    <tr>\n",
582 |        "      <th>Brackett, Jos</th>\n",
583 |        "      <td>0</td>\n",
584 |        "      <td>1</td>\n",
585 |        "      <td>0</td>\n",
586 |        "      <td>0</td>\n",
587 |        "      <td>0</td>\n",
588 |        "      <td>0</td>\n",
589 |        "      <td>0</td>\n",
590 |        "    </tr>\n",
591 |        "    <tr>\n",
592 |        "      <th>Bradford, John</th>\n",
593 |        "      <td>1</td>\n",
594 |        "      <td>1</td>\n",
595 |        "      <td>0</td>\n",
596 |        "      <td>0</td>\n",
597 |        "      <td>0</td>\n",
598 |        "      <td>0</td>\n",
599 |        "      <td>0</td>\n",
600 |        "    </tr>\n",
601 |        "  </tbody>\n",
602 |        "</table>\n",
603 |        "</div>"
604 |       ],
605 |       "text/plain": [
606 |        "cluster                                   0                             \\\n",
607 |        "org                         BostonCommittee LondonEnemies LongRoomClub   \n",
608 |        "cluster person                                                           \n",
609 |        "0       Adams, John                       0             0            1   \n",
610 |        "        Adams, Samuel                     1             1            1   \n",
611 |        "        Appleton, Nathaniel               1             0            0   \n",
612 |        "        Austin, Benjamin                  0             1            0   \n",
613 |        "        Austin, Samuel                    0             1            0   \n",
614 |        "        Baldwin, Cyrus                    0             1            0   \n",
615 |        "        Boyer, Peter                      0             1            0   \n",
616 |        "        Boynton, Richard                  1             1            0   \n",
617 |        "        Brackett, Jos                     0             1            0   \n",
618 |        "        Bradford, John                    1             1            0   \n",
619 |        "\n",
620 |        "cluster                            1              2           4         5  \n",
621 |        "org                         TeaParty StAndrewsLodge NorthCaucus LoyalNine  \n",
622 |        "cluster person                                                             \n",
623 |        "0       Adams, John                0              0           1         0  \n",
624 |        "        Adams, Samuel              0              0           1         0  \n",
625 |        "        Appleton, Nathaniel        0              0           1         0  \n",
626 |        "        Austin, Benjamin           0              0           0         0  \n",
627 |        "        Austin, Samuel             0              0           0         0  \n",
628 |        "        Baldwin, Cyrus             0              0           0         0  \n",
629 |        "        Boyer, Peter               0              0           0         0  \n",
630 |        "        Boynton, Richard           0              0           0         0  \n",
631 |        "        Brackett, Jos              0              0           0         0  \n",
632 |        "        Bradford, John             0              0           0         0  "
633 |       ]
634 |      },
635 |      "execution_count": 11,
636 |      "metadata": {},
637 |      "output_type": "execute_result"
638 |     }
639 |    ],
640 |    "source": [
641 |     "data_mi.head(10)"
642 |    ]
643 |   },
644 |   {
645 |    "cell_type": "markdown",
646 |    "metadata": {},
647 |    "source": [
648 |     "## Verify Coclustering structure\n",
649 |     "\"The resulting bicluster structure is block-diagonal, since each row and each column belongs to exactly one bicluster.\" - [sklearn SpectralCoclustering Docs](http://scikit-learn.org/stable/modules/generated/sklearn.cluster.bicluster.SpectralCoclustering.html)"
650 |    ]
651 |   },
652 |   {
653 |    "cell_type": "code",
654 |    "execution_count": 12,
655 |    "metadata": {},
656 |    "outputs": [
657 |     {
658 |      "data": {
659 |       "image/png": "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\n",
660 |       "text/plain": [
661 |        "<Figure size 1152x144 with 1 Axes>"
662 |       ]
663 |      },
664 |      "metadata": {
665 |       "needs_background": "light"
666 |      },
667 |      "output_type": "display_data"
668 |     }
669 |    ],
670 |    "source": [
671 |     "plt.matshow(data_mi.values.T, cmap=plt.cm.Blues, aspect=\"auto\", origin=\"lower\")\n",
672 |     "plt.xlabel(\"row index\")\n",
673 |     "plt.ylabel(\"column index\")\n",
674 |     "plt.show()\n",
675 |     "# I plotted the transpose so it doesn't give us a long image"
676 |    ]
677 |   },
678 |   {
679 |    "cell_type": "markdown",
680 |    "metadata": {},
681 |    "source": [
682 |     "## Brief Analysis\n",
683 |     "Let's take a look and see where the actual Loyal Nine members ended up. The organization ended up in our `Cluster 5`, but how many members of that bicluster match?\n",
684 |     "\n",
685 |     "Loyal Nine [members](https://books.google.com/books?id=iNgNCgAAQBAJ&pg=PA26#v=onepage&q&f=false\n",
686 |     "):\n",
687 |     "\n",
688 |     "- John Avery, a distiller and club secretary;\n",
689 |     "- John Smith and Stephen Cleverly, both braziers;\n",
690 |     "- Thomas Crafts, a printer;\n",
691 |     "- Benjamin Edes, who along with John Gill produced the important Boston Gazette;\n",
692 |     "- Thomas Chase, a distiller;\n",
693 |     "- Joseph Field, a ship's captain;\n",
694 |     "- George Trott, a jeweler;\n",
695 |     "- Henry Bass, a merchant related to Samuel Adams"
696 |    ]
697 |   },
698 |   {
699 |    "cell_type": "code",
700 |    "execution_count": 13,
701 |    "metadata": {},
702 |    "outputs": [],
703 |    "source": [
704 |     "loyal_nine = [\n",
705 |     "    \"Avery, John\",\n",
706 |     "    \"Smith, John\",\n",
707 |     "    \"Cleverly, Stephen\",\n",
708 |     "    \"Crafts, Thomas\",\n",
709 |     "    \"Edes, Benjamin\",\n",
710 |     "    \"Chase, Thomas\",\n",
711 |     "    \"Field, Joseph\",\n",
712 |     "    \"Trott, George\",\n",
713 |     "    \"Bass, Henry\",\n",
714 |     "]"
715 |    ]
716 |   },
717 |   {
718 |    "cell_type": "code",
719 |    "execution_count": 14,
720 |    "metadata": {},
721 |    "outputs": [
722 |     {
723 |      "data": {
724 |       "text/html": [
725 |        "<div>\n",
726 |        "<style scoped>\n",
727 |        "    .dataframe tbody tr th:only-of-type {\n",
728 |        "        vertical-align: middle;\n",
729 |        "    }\n",
730 |        "\n",
731 |        "    .dataframe tbody tr th {\n",
732 |        "        vertical-align: top;\n",
733 |        "    }\n",
734 |        "\n",
735 |        "    .dataframe thead tr th {\n",
736 |        "        text-align: left;\n",
737 |        "    }\n",
738 |        "\n",
739 |        "    .dataframe thead tr:last-of-type th {\n",
740 |        "        text-align: right;\n",
741 |        "    }\n",
742 |        "</style>\n",
743 |        "<table border=\"1\" class=\"dataframe\">\n",
744 |        "  <thead>\n",
745 |        "    <tr>\n",
746 |        "      <th></th>\n",
747 |        "      <th>cluster</th>\n",
748 |        "      <th colspan=\"3\" halign=\"left\">0</th>\n",
749 |        "      <th>1</th>\n",
750 |        "      <th>2</th>\n",
751 |        "      <th>4</th>\n",
752 |        "      <th>5</th>\n",
753 |        "    </tr>\n",
754 |        "    <tr>\n",
755 |        "      <th></th>\n",
756 |        "      <th>org</th>\n",
757 |        "      <th>BostonCommittee</th>\n",
758 |        "      <th>LondonEnemies</th>\n",
759 |        "      <th>LongRoomClub</th>\n",
760 |        "      <th>TeaParty</th>\n",
761 |        "      <th>StAndrewsLodge</th>\n",
762 |        "      <th>NorthCaucus</th>\n",
763 |        "      <th>LoyalNine</th>\n",
764 |        "    </tr>\n",
765 |        "    <tr>\n",
766 |        "      <th>cluster</th>\n",
767 |        "      <th>person</th>\n",
768 |        "      <th></th>\n",
769 |        "      <th></th>\n",
770 |        "      <th></th>\n",
771 |        "      <th></th>\n",
772 |        "      <th></th>\n",
773 |        "      <th></th>\n",
774 |        "      <th></th>\n",
775 |        "    </tr>\n",
776 |        "  </thead>\n",
777 |        "  <tbody>\n",
778 |        "    <tr>\n",
779 |        "      <th rowspan=\"2\" valign=\"top\">3</th>\n",
780 |        "      <th>Bass, Henry</th>\n",
781 |        "      <td>0</td>\n",
782 |        "      <td>1</td>\n",
783 |        "      <td>0</td>\n",
784 |        "      <td>1</td>\n",
785 |        "      <td>0</td>\n",
786 |        "      <td>1</td>\n",
787 |        "      <td>1</td>\n",
788 |        "    </tr>\n",
789 |        "    <tr>\n",
790 |        "      <th>Chase, Thomas</th>\n",
791 |        "      <td>0</td>\n",
792 |        "      <td>1</td>\n",
793 |        "      <td>0</td>\n",
794 |        "      <td>1</td>\n",
795 |        "      <td>0</td>\n",
796 |        "      <td>1</td>\n",
797 |        "      <td>1</td>\n",
798 |        "    </tr>\n",
799 |        "    <tr>\n",
800 |        "      <th rowspan=\"2\" valign=\"top\">4</th>\n",
801 |        "      <th>Avery, John</th>\n",
802 |        "      <td>0</td>\n",
803 |        "      <td>1</td>\n",
804 |        "      <td>0</td>\n",
805 |        "      <td>0</td>\n",
806 |        "      <td>0</td>\n",
807 |        "      <td>0</td>\n",
808 |        "      <td>1</td>\n",
809 |        "    </tr>\n",
810 |        "    <tr>\n",
811 |        "      <th>Edes, Benjamin</th>\n",
812 |        "      <td>0</td>\n",
813 |        "      <td>0</td>\n",
814 |        "      <td>0</td>\n",
815 |        "      <td>0</td>\n",
816 |        "      <td>0</td>\n",
817 |        "      <td>1</td>\n",
818 |        "      <td>1</td>\n",
819 |        "    </tr>\n",
820 |        "    <tr>\n",
821 |        "      <th rowspan=\"5\" valign=\"top\">5</th>\n",
822 |        "      <th>Cleverly, Stephen</th>\n",
823 |        "      <td>0</td>\n",
824 |        "      <td>0</td>\n",
825 |        "      <td>0</td>\n",
826 |        "      <td>0</td>\n",
827 |        "      <td>0</td>\n",
828 |        "      <td>0</td>\n",
829 |        "      <td>1</td>\n",
830 |        "    </tr>\n",
831 |        "    <tr>\n",
832 |        "      <th>Crafts, Thomas</th>\n",
833 |        "      <td>0</td>\n",
834 |        "      <td>0</td>\n",
835 |        "      <td>0</td>\n",
836 |        "      <td>0</td>\n",
837 |        "      <td>1</td>\n",
838 |        "      <td>0</td>\n",
839 |        "      <td>1</td>\n",
840 |        "    </tr>\n",
841 |        "    <tr>\n",
842 |        "      <th>Field, Joseph</th>\n",
843 |        "      <td>0</td>\n",
844 |        "      <td>0</td>\n",
845 |        "      <td>0</td>\n",
846 |        "      <td>0</td>\n",
847 |        "      <td>0</td>\n",
848 |        "      <td>0</td>\n",
849 |        "      <td>1</td>\n",
850 |        "    </tr>\n",
851 |        "    <tr>\n",
852 |        "      <th>Smith, John</th>\n",
853 |        "      <td>0</td>\n",
854 |        "      <td>0</td>\n",
855 |        "      <td>0</td>\n",
856 |        "      <td>0</td>\n",
857 |        "      <td>0</td>\n",
858 |        "      <td>0</td>\n",
859 |        "      <td>1</td>\n",
860 |        "    </tr>\n",
861 |        "    <tr>\n",
862 |        "      <th>Trott, George</th>\n",
863 |        "      <td>0</td>\n",
864 |        "      <td>0</td>\n",
865 |        "      <td>0</td>\n",
866 |        "      <td>0</td>\n",
867 |        "      <td>0</td>\n",
868 |        "      <td>0</td>\n",
869 |        "      <td>1</td>\n",
870 |        "    </tr>\n",
871 |        "  </tbody>\n",
872 |        "</table>\n",
873 |        "</div>"
874 |       ],
875 |       "text/plain": [
876 |        "cluster                                 0                                   1  \\\n",
877 |        "org                       BostonCommittee LondonEnemies LongRoomClub TeaParty   \n",
878 |        "cluster person                                                                  \n",
879 |        "3       Bass, Henry                     0             1            0        1   \n",
880 |        "        Chase, Thomas                   0             1            0        1   \n",
881 |        "4       Avery, John                     0             1            0        0   \n",
882 |        "        Edes, Benjamin                  0             0            0        0   \n",
883 |        "5       Cleverly, Stephen               0             0            0        0   \n",
884 |        "        Crafts, Thomas                  0             0            0        0   \n",
885 |        "        Field, Joseph                   0             0            0        0   \n",
886 |        "        Smith, John                     0             0            0        0   \n",
887 |        "        Trott, George                   0             0            0        0   \n",
888 |        "\n",
889 |        "cluster                                2           4         5  \n",
890 |        "org                       StAndrewsLodge NorthCaucus LoyalNine  \n",
891 |        "cluster person                                                  \n",
892 |        "3       Bass, Henry                    0           1         1  \n",
893 |        "        Chase, Thomas                  0           1         1  \n",
894 |        "4       Avery, John                    0           0         1  \n",
895 |        "        Edes, Benjamin                 0           1         1  \n",
896 |        "5       Cleverly, Stephen              0           0         1  \n",
897 |        "        Crafts, Thomas                 1           0         1  \n",
898 |        "        Field, Joseph                  0           0         1  \n",
899 |        "        Smith, John                    0           0         1  \n",
900 |        "        Trott, George                  0           0         1  "
901 |       ]
902 |      },
903 |      "execution_count": 14,
904 |      "metadata": {},
905 |      "output_type": "execute_result"
906 |     }
907 |    ],
908 |    "source": [
909 |     "data_mi[data_mi.index.get_level_values(\"person\").isin(loyal_nine)]"
910 |    ]
911 |   },
912 |   {
913 |    "cell_type": "markdown",
914 |    "metadata": {},
915 |    "source": [
916 |     "In `Cluster 5` we recovered all individuals who were *exclusively* members of the Loyal Nine, plus Thomas Crafts. `Cluster 4` pulled two Loyal Nine members also part of the North Caucus and London Enemies, and `Cluster 3` holds members that were additionally part of the North Cacus, Tea Party, and London Enemies each. Also note Cluster 3 contains no organizations -- it's members are part of several organizations."
917 |    ]
918 |   },
919 |   {
920 |    "cell_type": "markdown",
921 |    "metadata": {},
922 |    "source": [
923 |     "## Export to Excel\n",
924 |     "It's helpful to export this to Excel because the MultiIndex formats nicely into merged cells and retains the hierarchical indexing structure.\n",
925 |     "\n",
926 |     "The data is on the `data` sheet, and cluster details are on the `clusters` sheet."
927 |    ]
928 |   },
929 |   {
930 |    "cell_type": "code",
931 |    "execution_count": 15,
932 |    "metadata": {},
933 |    "outputs": [],
934 |    "source": [
935 |     "data_sheets = [(data_mi, \"data\"), (cluster_info, \"clusters\")]\n",
936 |     "\n",
937 |     "filename = \"bicluster.xlsx\"\n",
938 |     "writer = pd.ExcelWriter(filename, engine=\"xlsxwriter\")\n",
939 |     "for data in data_sheets:\n",
940 |     "    data[0].to_excel(writer, sheet_name=data[1])\n",
941 |     "writer.save()"
942 |    ]
943 |   },
944 |   {
945 |    "cell_type": "code",
946 |    "execution_count": null,
947 |    "metadata": {},
948 |    "outputs": [],
949 |    "source": []
950 |   }
951 |  ],
952 |  "metadata": {
953 |   "anaconda-cloud": {},
954 |   "kernelspec": {
955 |    "display_name": "Python 3",
956 |    "language": "python",
957 |    "name": "python3"
958 |   },
959 |   "language_info": {
960 |    "codemirror_mode": {
961 |     "name": "ipython",
962 |     "version": 3
963 |    },
964 |    "file_extension": ".py",
965 |    "mimetype": "text/x-python",
966 |    "name": "python",
967 |    "nbconvert_exporter": "python",
968 |    "pygments_lexer": "ipython3",
969 |    "version": "3.7.3"
970 |   }
971 |  },
972 |  "nbformat": 4,
973 |  "nbformat_minor": 1
974 | }
975 | 


--------------------------------------------------------------------------------
/binder/postBuild:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 | python -m spacy download en
3 | python -m spacy download en_core_web_md
4 | python -m gensim.downloader --download word2vec-google-news-300
5 | python -m nltk.downloader wordnet
6 | jupyter nbextension enable --py widgetsnbextension


--------------------------------------------------------------------------------
/binder/requirements.txt:
--------------------------------------------------------------------------------
 1 | altair==4.1.0
 2 | black==19.3b0
 3 | gensim==3.8.3
 4 | hdbscan==0.8.26
 5 | importlib-metadata==1.7.0
 6 | ipykernel==5.3.2
 7 | ipython==8.10.0
 8 | jsonschema==3.2.0
 9 | jupyter-client==6.1.5
10 | jupyter-core==4.6.3
11 | matplotlib==3.2.2
12 | nbformat>=4.2.0
13 | pandas==1.0.5
14 | plac==1.1.3
15 | plotly==4.14.3
16 | scikit-learn==0.23.1
17 | sentence-transformers==0.4.1.2
18 | spacy==2.3.1
19 | srsly==1.0.2
20 | tabulate==0.8.7
21 | tensorboard==2.2.2
22 | tensorboard-plugin-wit==1.7.0
23 | tensorflow==2.9.3
24 | tensorflow-estimator==2.2.0
25 | tensorflow-hub==0.8.0
26 | tqdm==4.47.0
27 | umap-learn==0.5
28 | 


--------------------------------------------------------------------------------
/cleaned_hm.csv.xz:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/pmbaumgartner/binder-notebooks/cfee899187c9ba5e7a11a5b26fb37aa78cbc1201/cleaned_hm.csv.xz


--------------------------------------------------------------------------------
/colored-roc-curves.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "markdown",
  5 |    "metadata": {},
  6 |    "source": [
  7 |     "# Colored ROC Curves"
  8 |    ]
  9 |   },
 10 |   {
 11 |    "cell_type": "markdown",
 12 |    "metadata": {},
 13 |    "source": [
 14 |     "The Receiver Operating Characteristic (ROC) curve is helpful in evaluating model performance, especially since Area Under the Curve (AUC ROC) has a several friendly interpretations. I use ROC curves in evaluating models and have to explain the output to non-technical folks. I was reading through [Machine Learning: The Art and Science of Algorithms that Make Sense of Data](http://a.co/dnKG3eO) and stumbled upon this nice visual and interpretation of ROC (tied to AUC):\n",
 15 |     "\n",
 16 |     "![png](https://i.imgur.com/nWXd3Q0.png)\n",
 17 |     "\n",
 18 |     ">Flach, P. (2012). Machine Learning: The Art and Science of Algorithms that Make Sense of Data. Cambridge University Press.\n",
 19 |     "\n",
 20 |     "The interpretation that each cell is a pair of true and false outcomes and their scores doesn't always map to reality -- usually you have more than 100 pairs and in imbalanced dataset -- but in general it helps me imagine model performance and explain to someone what the curve means."
 21 |    ]
 22 |   },
 23 |   {
 24 |    "cell_type": "code",
 25 |    "execution_count": 1,
 26 |    "metadata": {},
 27 |    "outputs": [],
 28 |    "source": [
 29 |     "import numpy as np\n",
 30 |     "import matplotlib.pyplot as plt\n",
 31 |     "\n",
 32 |     "from sklearn.datasets import load_digits\n",
 33 |     "from sklearn.tree import DecisionTreeClassifier\n",
 34 |     "from sklearn.metrics import roc_curve, auc\n",
 35 |     "from sklearn.model_selection import train_test_split\n",
 36 |     "\n",
 37 |     "%matplotlib inline"
 38 |    ]
 39 |   },
 40 |   {
 41 |    "cell_type": "markdown",
 42 |    "metadata": {},
 43 |    "source": [
 44 |     "## Load Data\n",
 45 |     "From the digits dataset we'll try to predict if a digit is `4` "
 46 |    ]
 47 |   },
 48 |   {
 49 |    "cell_type": "code",
 50 |    "execution_count": 2,
 51 |    "metadata": {},
 52 |    "outputs": [],
 53 |    "source": [
 54 |     "digits = load_digits()\n",
 55 |     "X = digits.data\n",
 56 |     "y = digits.target\n",
 57 |     "\n",
 58 |     "y = y == 4"
 59 |    ]
 60 |   },
 61 |   {
 62 |    "cell_type": "markdown",
 63 |    "metadata": {},
 64 |    "source": [
 65 |     "## Build Model\n",
 66 |     "We're making a purposefully bad model so the ROC curve is more interesting. Then building the ROC curve from predictions on a test set."
 67 |    ]
 68 |   },
 69 |   {
 70 |    "cell_type": "code",
 71 |    "execution_count": 3,
 72 |    "metadata": {},
 73 |    "outputs": [],
 74 |    "source": [
 75 |     "dt = DecisionTreeClassifier(max_depth=5)\n",
 76 |     "\n",
 77 |     "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=666)\n",
 78 |     "\n",
 79 |     "probas_ = dt.fit(X_train, y_train).predict_proba(X_test)\n",
 80 |     "\n",
 81 |     "fpr, tpr, thresholds = roc_curve(y_test, probas_[:, 1])\n",
 82 |     "roc_auc = auc(fpr, tpr)"
 83 |    ]
 84 |   },
 85 |   {
 86 |    "cell_type": "markdown",
 87 |    "metadata": {},
 88 |    "source": [
 89 |     "## Build Coloring Matrix\n",
 90 |     "We'll create a `100x100` matrix and color each chunk of cells according to the coordinates given by `FPR` and `TPR`"
 91 |    ]
 92 |   },
 93 |   {
 94 |    "cell_type": "code",
 95 |    "execution_count": 4,
 96 |    "metadata": {},
 97 |    "outputs": [],
 98 |    "source": [
 99 |     "coordlist = list(zip(((fpr * 100).astype(int)), ((tpr * 100).astype(int))))\n",
100 |     "\n",
101 |     "roc_array = np.zeros((100,100))\n",
102 |     "\n",
103 |     "for i, coords in enumerate(coordlist):\n",
104 |     "    if i == len(coordlist) - 1:\n",
105 |     "        continue\n",
106 |     "    else:\n",
107 |     "        # orange\n",
108 |     "        roc_array[coords[0]:coordlist[i+1][0], coords[1]:coordlist[i+1][1]] = 0\n",
109 |     "        # red\n",
110 |     "        roc_array[coords[0]:coordlist[i+1][0], coordlist[i+1][1]:100] = -1\n",
111 |     "        # green\n",
112 |     "        roc_array[coords[0]:100, 0:coords[1]] = 1"
113 |    ]
114 |   },
115 |   {
116 |    "cell_type": "markdown",
117 |    "metadata": {},
118 |    "source": [
119 |     "## Plot Colored Matrix & ROC Curve\n",
120 |     "There is still a small issue of the colors not exactly lining up with the gridlines."
121 |    ]
122 |   },
123 |   {
124 |    "cell_type": "code",
125 |    "execution_count": 5,
126 |    "metadata": {},
127 |    "outputs": [
128 |     {
129 |      "data": {
130 |       "image/png": "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\n",
131 |       "text/plain": [
132 |        "<Figure size 576x576 with 1 Axes>"
133 |       ]
134 |      },
135 |      "metadata": {},
136 |      "output_type": "display_data"
137 |     }
138 |    ],
139 |    "source": [
140 |     "fig = plt.figure(figsize=(8,8))\n",
141 |     "\n",
142 |     "# ROC Curve\n",
143 |     "plt.plot((fpr * 100).astype(int), (tpr * 100).astype(int),\n",
144 |     "         lw=1.5, linestyle='--', color='navy',\n",
145 |     "         label='ROC (AUC = %0.2f)' % (roc_auc))\n",
146 |     "\n",
147 |     "# Dummy Model\n",
148 |     "plt.plot([0.0, 100.0], [0.0, 100.0],\n",
149 |     "         linestyle='-', lw=1, color='k', alpha=0.75,\n",
150 |     "         label='Baseline')\n",
151 |     "\n",
152 |     "# ROC Color Matrix\n",
153 |     "plt.imshow(roc_array.T, \n",
154 |     "           cmap=plt.cm.RdYlGn, alpha=0.9, origin='lower',\n",
155 |     "           aspect='auto', interpolation='nearest', extent=[0.0, 100.0, 0.0, 100.0])\n",
156 |     "plt.legend(loc=\"lower right\")\n",
157 |     "\n",
158 |     "# adding the lines to the chart automatically extends the axis by (-20, -20)\n",
159 |     "ax = fig.axes[0]\n",
160 |     "ax.set_xlim([0,100])\n",
161 |     "ax.set_ylim([0,100])\n",
162 |     "\n",
163 |     "# axis labels\n",
164 |     "plt.xlabel('Negatives sorted on decreasing score')\n",
165 |     "plt.ylabel('Positives sorted on decreasing score')\n",
166 |     "\n",
167 |     "# minor ticks each unit\n",
168 |     "minor_ticks = np.arange(0, 100, 1)                                               \n",
169 |     "ax.set_xticks(minor_ticks, minor=True)                                                                                                 \n",
170 |     "ax.set_yticks(minor_ticks, minor=True)\n",
171 |     "ax.grid(which='minor', alpha=0.6, linestyle='-')\n",
172 |     "\n",
173 |     "# tight_layout for figures sometimes getting cutoff\n",
174 |     "plt.tight_layout()\n",
175 |     "plt.savefig('roc_colored.png')"
176 |    ]
177 |   },
178 |   {
179 |    "cell_type": "markdown",
180 |    "metadata": {},
181 |    "source": [
182 |     "## Relevant `matplotlib` StackOverflow Posts\n",
183 |     "- [matching axis in imshow](http://stackoverflow.com/questions/12324176/matplotlib-imshow-offset-to-match-axis)\n",
184 |     "- [limiting y-axis](http://stackoverflow.com/questions/3777861/setting-y-axis-limit-in-matplotlib)\n",
185 |     "- [specifing minor axis grid aesthetics](http://stackoverflow.com/questions/24943991/matplotlib-change-grid-interval-and-specify-tick-labels)"
186 |    ]
187 |   }
188 |  ],
189 |  "metadata": {
190 |   "anaconda-cloud": {},
191 |   "kernelspec": {
192 |    "display_name": "Python 3",
193 |    "language": "python",
194 |    "name": "python3"
195 |   }
196 |  },
197 |  "nbformat": 4,
198 |  "nbformat_minor": 1
199 | }


--------------------------------------------------------------------------------
/nlp-sbert-umap-hdbscan.ipynb:
--------------------------------------------------------------------------------
1 | {"cells":[{"cell_type":"markdown","metadata":{},"source":["# Sentence Embeddings, Custom Distances, and Spherical Projections\n","This notebook is a workflow for generating, analyzing, and interpreting embeddings of sentences. If you're used to the basic workflow of getting embeddings and projecting them to two dimensions, this notebook contains some unique jumping off points:\n","\n","1. It uses the `paraphrase-distilroberta-base-v1` model from the `sentence-transformers` library instead of Universal Sentence Encoder. I find this gives much better results.\n","2. It uses a custom distance metric, TS-SS, from the paper: [A. Heidarian and M. J. Dinneen, \"A Hybrid Geometric Approach for Measuring Similarity Level Among Documents and Document Clustering,\"](https://ieeexplore.ieee.org/document/7474366?reload=true). I used [this implementation](https://github.com/taki0112/Vector_Similarity/blob/master/README.md) as a reference, but corrected an error and turned it into a numba function for faster computation.\n","3. In projecting using UMAP, it projects to onto spherical coordinates instead of a traditional Euclidean 2D plane. I personally believe this leads to a much more intuitive understanding of embeddings. Rather than an interpretation on a 2D plane where in some sense we are \"bound\" by the limits of the x-axis and y-axis, it makes more sense to me that the embedding space is \"wrapped\" and doesn't have an end.\n","  1. We are also able to use this spherical projection with haversine distance to develop clusters derived from the spherical projection rather than the planar projection."]},{"cell_type":"code","execution_count":1,"metadata":{},"outputs":[],"source":["import altair as alt\n","import numpy as np\n","import pandas as pd\n","import plotly.express as px\n","from hdbscan import HDBSCAN\n","from sentence_transformers import SentenceTransformer\n","from umap import UMAP\n","import numba\n"]},{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[],"source":["@numba.njit()\n","def tsss(vec1, vec2):\n","    euclidean_distance = np.linalg.norm(vec1 - vec2)\n","    cosine_distance = np.dot(vec1, vec2.T) / (np.linalg.norm(vec1) * np.linalg.norm(vec2))\n","    magnitude_difference = abs(np.linalg.norm(vec1) - np.linalg.norm(vec2))\n","    theta = np.arccos(cosine_distance) + np.radians(10.0)\n","    triangle = (np.linalg.norm(vec1) * np.linalg.norm(vec2) * np.sin(theta)) / 2\n","    sector = np.pi * ((euclidean_distance + magnitude_difference) ** 2) * (np.degrees(theta) / 360)\n","    return triangle * sector\n"]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[],"source":["df = (\n","    pd.read_csv(\"cleaned_hm.csv.xz\")\n","    .dropna(subset=[\"ground_truth_category\", \"cleaned_hm\"], how=\"any\")\n","    .drop_duplicates(subset=[\"cleaned_hm\"])\n","    .loc[:, [\"cleaned_hm\", \"ground_truth_category\"]]\n","    .reset_index(drop=True)\n",")\n"]},{"cell_type":"code","execution_count":4,"metadata":{},"outputs":[],"source":["model = SentenceTransformer(\"paraphrase-distilroberta-base-v1\")\n","\n"]},{"cell_type":"code","execution_count":5,"metadata":{},"outputs":[{"name":"stderr","output_type":"stream","text":["Batches: 100%|██████████| 199/199 [05:54<00:00,  1.78s/it]\n"]}],"source":["docs = df[\"cleaned_hm\"].str.lower().tolist()\n","embeddings = model.encode(docs, show_progress_bar=True, device=\"cpu\", batch_size=64)\n"]},{"cell_type":"code","execution_count":6,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["UMAP(dens_frac=0.0, dens_lambda=0.0,\n","     metric=CPUDispatcher(<function tsss at 0x7fd04b217290>), min_dist=0.0,\n","     n_neighbors=30, random_state=1234, verbose=True)\n","Construct fuzzy simplicial set\n","Fri Jan 29 08:42:42 2021 Finding Nearest Neighbors\n","Fri Jan 29 08:42:42 2021 Building RP forest with 11 trees\n","Fri Jan 29 08:42:43 2021 NN descent for 14 iterations\n","\t 1  /  14\n","\t 2  /  14\n","\t 3  /  14\n","\t 4  /  14\n","\tStopping threshold met -- exiting after 4 iterations\n","Fri Jan 29 08:44:31 2021 Finished Nearest Neighbor Search\n","Fri Jan 29 08:44:34 2021 Construct embedding\n","\tcompleted  0  /  200 epochs\n","\tcompleted  20  /  200 epochs\n","\tcompleted  40  /  200 epochs\n","\tcompleted  60  /  200 epochs\n","\tcompleted  80  /  200 epochs\n","\tcompleted  100  /  200 epochs\n","\tcompleted  120  /  200 epochs\n","\tcompleted  140  /  200 epochs\n","\tcompleted  160  /  200 epochs\n","\tcompleted  180  /  200 epochs\n","Fri Jan 29 08:44:44 2021 Finished embedding\n"]}],"source":["umap = UMAP(min_dist=0.00, n_neighbors=30, metric=tsss, random_state=1234, verbose=True)\n","embeddings_umap = umap.fit_transform(embeddings)\n"]},{"cell_type":"code","execution_count":7,"metadata":{},"outputs":[],"source":["embeddings_df = pd.DataFrame()\n","embeddings_df[\"Document\"] = docs\n","embeddings_df[\"Component 1\"] = embeddings_umap[:, 0]\n","embeddings_df[\"Component 2\"] = embeddings_umap[:, 1]\n","embeddings_df[\"Ground Truth\"] = df[\"ground_truth_category\"]\n"]},{"cell_type":"code","execution_count":8,"metadata":{},"outputs":[],"source":["hdbscan = HDBSCAN(cluster_selection_method=\"leaf\")\n","\n"]},{"cell_type":"code","execution_count":9,"metadata":{},"outputs":[],"source":["clusters = hdbscan.fit_predict(embeddings_umap)\n","embeddings_df[\"Cluster\"] = clusters\n"]},{"cell_type":"code","execution_count":10,"metadata":{},"outputs":[{"data":{"text/plain":"DataTransformerRegistry.enable('default')"},"execution_count":10,"metadata":{},"output_type":"execute_result"}],"source":["alt.data_transformers.disable_max_rows()\n"]},{"cell_type":"code","execution_count":11,"metadata":{},"outputs":[],"source":["chart = (\n","    (\n","        alt.Chart(\n","            embeddings_df,\n","            height=1000,\n","            width=1000,\n","            title=\"Happy Moments - SBERT → UMAP (TSSS) → HDBSCAN\",\n","        )\n","        .mark_point()\n","        .encode(\n","            x=alt.X(\"Component 1\", axis=None),\n","            y=alt.Y(\"Component 2\", axis=None),\n","            tooltip=[\"Document\", \"Cluster\", \"Ground Truth\"],\n","            color=\"Cluster:N\",\n","        )\n","    )\n","    .configure_axis(grid=False)\n","    .configure_view(strokeWidth=0)\n","    .interactive()\n",")\n"]},{"cell_type":"code","execution_count":12,"metadata":{},"outputs":[],"source":["chart.save(\"docs/nlp-sbert-umap-tsss-hdbscan-chart.html\")\n"]},{"cell_type":"code","execution_count":13,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["UMAP(dens_frac=0.0, dens_lambda=0.0,\n","     metric=CPUDispatcher(<function tsss at 0x7fd04b217290>), min_dist=0.0,\n","     n_neighbors=30, output_metric='haversine', random_state=42, verbose=True)\n","Construct fuzzy simplicial set\n","Fri Jan 29 08:44:44 2021 Finding Nearest Neighbors\n","Fri Jan 29 08:44:44 2021 Building RP forest with 11 trees\n","Fri Jan 29 08:44:45 2021 NN descent for 14 iterations\n","\t 1  /  14\n","\t 2  /  14\n","\t 3  /  14\n","\t 4  /  14\n","\tStopping threshold met -- exiting after 4 iterations\n","Fri Jan 29 08:46:24 2021 Finished Nearest Neighbor Search\n","Fri Jan 29 08:46:24 2021 Construct embedding\n","\tcompleted  0  /  200 epochs\n","\tcompleted  20  /  200 epochs\n","\tcompleted  40  /  200 epochs\n","\tcompleted  60  /  200 epochs\n","\tcompleted  80  /  200 epochs\n","\tcompleted  100  /  200 epochs\n","\tcompleted  120  /  200 epochs\n","\tcompleted  140  /  200 epochs\n","\tcompleted  160  /  200 epochs\n","\tcompleted  180  /  200 epochs\n","Fri Jan 29 08:47:33 2021 Finished embedding\n"]}],"source":["sphere_mapper = UMAP(\n","    min_dist=0.00,\n","    n_neighbors=30,\n","    metric=tsss,\n","    output_metric=\"haversine\",\n","    random_state=42,\n","    verbose=True,\n",")\n","umap_sphere_embeddings = sphere_mapper.fit_transform(embeddings)\n","\n","e1, e2 = umap_sphere_embeddings[:, 0], umap_sphere_embeddings[:, 1]\n","\n","x = np.sin(e1) * np.cos(e2)\n","y = np.sin(e1) * np.sin(e2)\n","z = np.cos(e1)\n","\n","embeddings_df[\"Sphere Embedding X\"] = x\n","embeddings_df[\"Sphere Embedding Y\"] = y\n","embeddings_df[\"Sphere Embedding Z\"] = z\n","\n","embeddings_df[\"2D Sphere Projection X\"] = np.arctan2(x, y)\n","embeddings_df[\"2D Sphere Projection Y\"] = -np.arccos(z)\n"]},{"cell_type":"code","execution_count":14,"metadata":{},"outputs":[],"source":["hdbscan_sphere = HDBSCAN(cluster_selection_method=\"eom\", metric=\"haversine\")\n","clusters_sphere = hdbscan_sphere.fit_predict(\n","    np.radians(embeddings_df[[\"2D Sphere Projection X\", \"2D Sphere Projection Y\"]].values)\n",")\n","embeddings_df[\"Cluster Sphere\"] = clusters_sphere\n","embeddings_df[\"Cluster Sphere (str)\"] = embeddings_df[\"Cluster Sphere\"].astype(str)\n"]},{"cell_type":"code","execution_count":15,"metadata":{},"outputs":[],"source":["fig = px.scatter(\n","    embeddings_df,\n","    x=\"2D Sphere Projection X\",\n","    y=\"2D Sphere Projection Y\",\n","    color=\"Cluster Sphere (str)\",\n","    hover_data=[\"Document\", \"Cluster Sphere\", \"Ground Truth\"],\n","    opacity=0.8,\n","    color_discrete_sequence=px.colors.qualitative.Alphabet,\n",")\n","\n","fig.write_html(\"docs/nlp-sbert-umap-hdbscan-tsss-sphere-2d.html\")\n"]},{"cell_type":"code","execution_count":16,"metadata":{},"outputs":[],"source":["fig = px.scatter_3d(\n","    embeddings_df,\n","    x=\"Sphere Embedding X\",\n","    y=\"Sphere Embedding Y\",\n","    z=\"Sphere Embedding Z\",\n","    color=\"Cluster Sphere (str)\",\n","    hover_data=[\"Document\", \"Cluster Sphere\", \"Ground Truth\"],\n","    color_discrete_sequence=px.colors.qualitative.Alphabet,\n","    opacity=0.8,\n",")\n","\n","camera = dict(up=dict(x=0, y=0, z=0), center=dict(x=0, y=0, z=0), eye=dict(x=0, y=0, z=0))\n","\n","fig.update_layout(\n","    scene_camera=camera,\n","    scene=dict(\n","        bgcolor=\"black\",\n","        xaxis=dict(\n","            showbackground=False,\n","            gridcolor=\"black\",\n","            zerolinecolor=\"black\",\n","        ),\n","        yaxis=dict(\n","            showbackground=False,\n","            gridcolor=\"black\",\n","            zerolinecolor=\"black\",\n","        ),\n","        zaxis=dict(\n","            showbackground=False,\n","            gridcolor=\"black\",\n","            zerolinecolor=\"black\",\n","        ),\n","    ),\n",")\n","\n","fig.write_html(\"docs/nlp-sbert-umap-tsss-hdbscan-sphere.html\")\n"]},{"cell_type":"code","execution_count":null,"metadata":{},"outputs":[],"source":[]}],"metadata":{"kernelspec":{"display_name":"Python 3.7.7 64-bit ('binder-notebooks': conda)","metadata":{"interpreter":{"hash":"71584da51981f72ec8f61ce2b9591b38d12385468c0adbe248ba9aa552e53946"}},"name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.7.7"},"orig_nbformat":2},"nbformat":4,"nbformat_minor":2}


--------------------------------------------------------------------------------
/nlp-use-umap-hdbscan.ipynb:
--------------------------------------------------------------------------------
1 | {"cells":[{"cell_type":"code","execution_count":1,"metadata":{},"outputs":[],"source":["from itertools import chain\n","\n","import altair as alt\n","import pandas as pd\n","import tensorflow_hub as hub\n","from hdbscan import HDBSCAN\n","from IPython.display import display\n","from umap import UMAP"]},{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[],"source":["df = (\n","    pd.read_csv(\"cleaned_hm.csv.xz\")\n","    .dropna(subset=[\"ground_truth_category\", \"cleaned_hm\"], how=\"any\")\n","    .loc[:, [\"cleaned_hm\", \"ground_truth_category\"]]\n","    .reset_index(drop=True)\n",")\n","\n",""]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[],"source":["embed = hub.load(\"https://tfhub.dev/google/universal-sentence-encoder/4\")"]},{"cell_type":"code","execution_count":4,"metadata":{},"outputs":[],"source":["docs = df[\"cleaned_hm\"].tolist()\n","embeddings = embed(docs)"]},{"cell_type":"code","execution_count":5,"metadata":{},"outputs":[],"source":["umap = UMAP(min_dist=0.00, n_neighbors=30, metric=\"cosine\", random_state=1234)\n","embeddings_umap = umap.fit_transform(embeddings)"]},{"cell_type":"code","execution_count":6,"metadata":{},"outputs":[],"source":["embeddings_df = pd.DataFrame()\n","embeddings_df[\"Document\"] = docs\n","embeddings_df[\"Component 1\"] = embeddings_umap[:, 0]\n","embeddings_df[\"Component 2\"] = embeddings_umap[:, 1]\n","embeddings_df[\"Ground Truth\"] = df[\"ground_truth_category\"]"]},{"cell_type":"code","execution_count":7,"metadata":{},"outputs":[],"source":["hdbscan = HDBSCAN(cluster_selection_method=\"leaf\", min_cluster_size=len(embeddings_df) // 100)"]},{"cell_type":"code","execution_count":8,"metadata":{},"outputs":[],"source":["clusters = hdbscan.fit_predict(embeddings_umap)\n","embeddings_df[\"Cluster\"] = clusters"]},{"cell_type":"code","execution_count":9,"metadata":{},"outputs":[{"output_type":"execute_result","data":{"text/plain":"DataTransformerRegistry.enable('default')"},"metadata":{},"execution_count":9}],"source":["alt.data_transformers.disable_max_rows()"]},{"cell_type":"code","execution_count":10,"metadata":{},"outputs":[],"source":["chart = (\n","    (\n","        alt.Chart(\n","            embeddings_df,\n","            height=1000,\n","            width=1000,\n","            title=\"Happy Moments - USE → UMAP → HDBSCAN\",\n","        )\n","        .mark_point()\n","        .encode(\n","            x=alt.X(\"Component 1\", axis=None),\n","            y=alt.Y(\"Component 2\", axis=None),\n","            tooltip=[\"Document\", \"Cluster\", \"Ground Truth\"],\n","            color=\"Cluster:N\",\n","            shape=\"Ground Truth\"\n","        )\n","    )\n","    .configure_axis(grid=False)\n","    .configure_view(strokeWidth=0)\n","    .interactive()\n",")"]},{"cell_type":"code","execution_count":11,"metadata":{},"outputs":[],"source":["chart.save(\"docs/nlp-use-umap-hdbscan-chart.html\")"]},{"cell_type":"code","execution_count":12,"metadata":{},"outputs":[],"source":["exemplars = list(map(tuple, chain.from_iterable(hdbscan.exemplars_)))\n","\n","def exemplar(row):\n","    return (row[\"Component 1\"], row[\"Component 2\"]) in exemplars\n","\n","embeddings_df[\"Exemplar\"] = embeddings_df.apply(exemplar, axis=1)\n"]},{"cell_type":"code","execution_count":13,"metadata":{},"outputs":[{"output_type":"execute_result","data":{"text/plain":"False    10829\nTrue      3296\nName: Exemplar, dtype: int64"},"metadata":{},"execution_count":13}],"source":["embeddings_df[\"Exemplar\"].value_counts()"]},{"cell_type":"code","execution_count":14,"metadata":{},"outputs":[],"source":["grouped_df = embeddings_df.groupby(\"Cluster\")"]},{"cell_type":"code","execution_count":15,"metadata":{},"outputs":[],"source":["pd.set_option(\"display.max_colwidth\", 200)"]},{"cell_type":"code","execution_count":18,"metadata":{"tags":[]},"outputs":[{"output_type":"stream","name":"stdout","text":"No Cluster (outlier) \tN Total: 6831 (48.4%)\n-------------------------\nCluster: 0 \tN Total: 144 (1.0%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                       Document\n606             I WENT TO MOVIE\n749             I WEND TO MOVIE\n750             I WENT TO MOVIE\n875             I WENT TO MOVIE\n877             I WENT TO MOVIE\n...                         ...\n13909           I WENT TO MOVIE\n13992              I WENT MOVIE\n13998           I WENT TO MOVIE\n14069           I WENT TO MOVIE\n14096  I saw a romantic comedy.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>606</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>749</th>\n      <td>I WEND TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>750</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>875</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>877</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13909</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>13992</th>\n      <td>I WENT MOVIE</td>\n    </tr>\n    <tr>\n      <th>13998</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>14069</th>\n      <td>I WENT TO MOVIE</td>\n    </tr>\n    <tr>\n      <th>14096</th>\n      <td>I saw a romantic comedy.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"         Ground Truth\nleisure           144","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>leisure</th>\n      <td>144</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 1 \tN Total: 153 (1.1%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                    Document\n94                                                                   My cat snuggled with me while we slept.\n241                                                            My cat cuddled with me when I went to sleep. \n256                                        I was able to take my cat out for a walk while the sky was clear.\n257                                           I played with my new kitten, we have only had him for 3 weeks!\n272                                                                   My friend sent me pictures of his cat.\n...                                                                                                      ...\n13655                        We found a stray cat and couldnt catch it but we fed it . that made me happy . \n13666                  My wife found a kitten on her morning walk.  She is really cute and my kids love her.\n13754       My cats came and cuddled with me, including the male the cat that does not like the cuddle much.\n13791                                                                    My son learned to meow like a cat. \n13808  An event that made me happy in the past 24 hours was petting my cats after feeding them this morning.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>94</th>\n      <td>My cat snuggled with me while we slept.</td>\n    </tr>\n    <tr>\n      <th>241</th>\n      <td>My cat cuddled with me when I went to sleep.</td>\n    </tr>\n    <tr>\n      <th>256</th>\n      <td>I was able to take my cat out for a walk while the sky was clear.</td>\n    </tr>\n    <tr>\n      <th>257</th>\n      <td>I played with my new kitten, we have only had him for 3 weeks!</td>\n    </tr>\n    <tr>\n      <th>272</th>\n      <td>My friend sent me pictures of his cat.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13655</th>\n      <td>We found a stray cat and couldnt catch it but we fed it . that made me happy .</td>\n    </tr>\n    <tr>\n      <th>13666</th>\n      <td>My wife found a kitten on her morning walk.  She is really cute and my kids love her.</td>\n    </tr>\n    <tr>\n      <th>13754</th>\n      <td>My cats came and cuddled with me, including the male the cat that does not like the cuddle much.</td>\n    </tr>\n    <tr>\n      <th>13791</th>\n      <td>My son learned to meow like a cat.</td>\n    </tr>\n    <tr>\n      <th>13808</th>\n      <td>An event that made me happy in the past 24 hours was petting my cats after feeding them this morning.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  124\nenjoy_the_moment            11\nachievement                  8\nbonding                      5\nleisure                      5","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>124</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>11</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>8</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>5</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>5</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 2 \tN Total: 159 (1.1%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                     Document\n56                                                                                                                                                       I bought a new phone\n97                                                                                                                                                     I bought a new laptop.\n171                                         I picked up a new phone this afternoon with an Android operating system instead of the terrible Windows phone I have been using. \n337                                                                                                        I bought a new cell phone with latest features and specifications.\n401                                                                                                                           I've got a new laptop and it works really well.\n...                                                                                                                                                                       ...\n13545                                                                                                                                                  I bought a new laptop.\n13683                                                                                                                    I managed to install a new screen on a broken phone.\n13903                                                                                                                                                  I bought a new laptop.\n13959                           I was shopping for a new cell phone at Target and found out that I will receive a $200 gift card when I buy a new phone and activate service.\n14107  My smart phone was giving problems due to virus and I repaired my phone with a help of Internet videos. I am very happy with that because I am not a service engineer.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>56</th>\n      <td>I bought a new phone</td>\n    </tr>\n    <tr>\n      <th>97</th>\n      <td>I bought a new laptop.</td>\n    </tr>\n    <tr>\n      <th>171</th>\n      <td>I picked up a new phone this afternoon with an Android operating system instead of the terrible Windows phone I have been using.</td>\n    </tr>\n    <tr>\n      <th>337</th>\n      <td>I bought a new cell phone with latest features and specifications.</td>\n    </tr>\n    <tr>\n      <th>401</th>\n      <td>I've got a new laptop and it works really well.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13545</th>\n      <td>I bought a new laptop.</td>\n    </tr>\n    <tr>\n      <th>13683</th>\n      <td>I managed to install a new screen on a broken phone.</td>\n    </tr>\n    <tr>\n      <th>13903</th>\n      <td>I bought a new laptop.</td>\n    </tr>\n    <tr>\n      <th>13959</th>\n      <td>I was shopping for a new cell phone at Target and found out that I will receive a $200 gift card when I buy a new phone and activate service.</td>\n    </tr>\n    <tr>\n      <th>14107</th>\n      <td>My smart phone was giving problems due to virus and I repaired my phone with a help of Internet videos. I am very happy with that because I am not a service engineer.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                134\naffection                    9\nenjoy_the_moment             8\nleisure                      5\nbonding                      3","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>134</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>9</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>8</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>5</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>3</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 3 \tN Total: 299 (2.1%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                            Document\n114                                                                                           My friend, Sheryl came over to visit me and my puppy. \n235                                                       Spending time with my dogs outside made me happy. I love watching them play and having fun\n285                                            Both of my dogs and my toddler played together nicely outside while I enjoyed the nice sunny weather.\n521                                                                            I petted and played with a friend's beautiful dog (golden retriever).\n532                             My dogs were all three vying for my attention by jumping at my feet and simultaneous giving me lots of doggy kisses.\n...                                                                                                                                              ...\n13488                                                                     I was able to play with my parents' dog for the first time in a few weeks.\n13775                                                                     I took my dog on a drive around town and she seemed really happy about it.\n13783  My fiancee and our two dogs were all lying in bed comfortably. I watched them just lie there for a couple minutes before going to bed myself.\n13854           I was happy when my dog showed improvement with a new trick that I am trying to teach her when I worked on it with her this morning.\n14097        I was happy tonight coming home and seeing my dog wag his tail and give me big lick on the cheek after missing me while I was at work. \n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>114</th>\n      <td>My friend, Sheryl came over to visit me and my puppy.</td>\n    </tr>\n    <tr>\n      <th>235</th>\n      <td>Spending time with my dogs outside made me happy. I love watching them play and having fun</td>\n    </tr>\n    <tr>\n      <th>285</th>\n      <td>Both of my dogs and my toddler played together nicely outside while I enjoyed the nice sunny weather.</td>\n    </tr>\n    <tr>\n      <th>521</th>\n      <td>I petted and played with a friend's beautiful dog (golden retriever).</td>\n    </tr>\n    <tr>\n      <th>532</th>\n      <td>My dogs were all three vying for my attention by jumping at my feet and simultaneous giving me lots of doggy kisses.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13488</th>\n      <td>I was able to play with my parents' dog for the first time in a few weeks.</td>\n    </tr>\n    <tr>\n      <th>13775</th>\n      <td>I took my dog on a drive around town and she seemed really happy about it.</td>\n    </tr>\n    <tr>\n      <th>13783</th>\n      <td>My fiancee and our two dogs were all lying in bed comfortably. I watched them just lie there for a couple minutes before going to bed myself.</td>\n    </tr>\n    <tr>\n      <th>13854</th>\n      <td>I was happy when my dog showed improvement with a new trick that I am trying to teach her when I worked on it with her this morning.</td>\n    </tr>\n    <tr>\n      <th>14097</th>\n      <td>I was happy tonight coming home and seeing my dog wag his tail and give me big lick on the cheek after missing me while I was at work.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  246\nachievement                 17\nbonding                     17\nenjoy_the_moment             6\nleisure                      5\nnature                       4\nexercise                     4","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>246</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>17</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>17</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>6</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>5</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>4</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>4</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 4 \tN Total: 222 (1.6%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n109                                                                                                                  I watched a wonderful movie that I have not seen in years. My family and I were thrilled.\n138                                                                                                                                                    I watched a movie with my wife, and we both enjoyed it.\n430                                                                                                                         I was able to see one of my most anticipated movies, Logan, starring Hugh Jackman.\n540                                                                                                                    In the last 24 hours I was happy because I went to a movie theater and enjoyed a movie.\n696                                                                                                                                  There was a funny scene in a movie that I was seeing for the first time. \n...                                                                                                                                                                                                        ...\n13425                                                                                                       today made me happy to invite my wife and my best childhood friend to watch a movie in the cinema.\n13682                                                                                                                 My sister and I watched \"An Affair to Remember\" and really enjoyed watching it together.\n13773                                                                                                                                                          I love watch movies with my wife in my bedroom \n13809                                                                                                                                                                   watched Bahubali movie with my friends\n13957  An event that made me happy is my lyrics as be selected for a movie. When my lyrics is played on the big screen it brings me proud movement. The Director and the Hero said if this lyrics is played...\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>109</th>\n      <td>I watched a wonderful movie that I have not seen in years. My family and I were thrilled.</td>\n    </tr>\n    <tr>\n      <th>138</th>\n      <td>I watched a movie with my wife, and we both enjoyed it.</td>\n    </tr>\n    <tr>\n      <th>430</th>\n      <td>I was able to see one of my most anticipated movies, Logan, starring Hugh Jackman.</td>\n    </tr>\n    <tr>\n      <th>540</th>\n      <td>In the last 24 hours I was happy because I went to a movie theater and enjoyed a movie.</td>\n    </tr>\n    <tr>\n      <th>696</th>\n      <td>There was a funny scene in a movie that I was seeing for the first time.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13425</th>\n      <td>today made me happy to invite my wife and my best childhood friend to watch a movie in the cinema.</td>\n    </tr>\n    <tr>\n      <th>13682</th>\n      <td>My sister and I watched \"An Affair to Remember\" and really enjoyed watching it together.</td>\n    </tr>\n    <tr>\n      <th>13773</th>\n      <td>I love watch movies with my wife in my bedroom</td>\n    </tr>\n    <tr>\n      <th>13809</th>\n      <td>watched Bahubali movie with my friends</td>\n    </tr>\n    <tr>\n      <th>13957</th>\n      <td>An event that made me happy is my lyrics as be selected for a movie. When my lyrics is played on the big screen it brings me proud movement. The Director and the Hero said if this lyrics is played...</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nleisure                     96\naffection                   70\nbonding                     37\nenjoy_the_moment            15\nachievement                  4","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>leisure</th>\n      <td>96</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>70</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>37</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>15</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>4</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 5 \tN Total: 179 (1.3%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                Document\n51                                                                         Watched a new episode of a show that I like. \n99                                                                              Watching Bob's Burgers streaming on Fox.\n131                                       I watched a particularly compelling episode of a show that I enjoy on Netflix.\n219                                   I saw that there was a new episode of Mindy Project and I was excited to watch it.\n228                                                                                    I watched a TV show I like a lot.\n...                                                                                                                  ...\n13803                           My favorite TV show is back on for a new season on AMC, and I can't wait to binge on it.\n13851                                                              I managed to find time to catch up on Netflix series.\n13885                                      I was able to watch my favorite show, which I hadn't had time for in a while.\n13974                                                                              i watched the game of thrones trailer\n14078  I have a favorite tv show, and today a new episode which I did not realize was going to air, happened to come on.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>51</th>\n      <td>Watched a new episode of a show that I like.</td>\n    </tr>\n    <tr>\n      <th>99</th>\n      <td>Watching Bob's Burgers streaming on Fox.</td>\n    </tr>\n    <tr>\n      <th>131</th>\n      <td>I watched a particularly compelling episode of a show that I enjoy on Netflix.</td>\n    </tr>\n    <tr>\n      <th>219</th>\n      <td>I saw that there was a new episode of Mindy Project and I was excited to watch it.</td>\n    </tr>\n    <tr>\n      <th>228</th>\n      <td>I watched a TV show I like a lot.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13803</th>\n      <td>My favorite TV show is back on for a new season on AMC, and I can't wait to binge on it.</td>\n    </tr>\n    <tr>\n      <th>13851</th>\n      <td>I managed to find time to catch up on Netflix series.</td>\n    </tr>\n    <tr>\n      <th>13885</th>\n      <td>I was able to watch my favorite show, which I hadn't had time for in a while.</td>\n    </tr>\n    <tr>\n      <th>13974</th>\n      <td>i watched the game of thrones trailer</td>\n    </tr>\n    <tr>\n      <th>14078</th>\n      <td>I have a favorite tv show, and today a new episode which I did not realize was going to air, happened to come on.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nleisure                    146\nenjoy_the_moment            13\naffection                   13\nbonding                      4\nachievement                  3","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>leisure</th>\n      <td>146</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>13</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>13</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>4</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>3</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 6 \tN Total: 309 (2.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                     Document\n122                                                                                      Increased my max bench press this morning, heaviest I've ever lifted\n147                                             I went to the gym. Hitting a new personal record on the bench press made me more confident in my abilities.  \n214                                                                    I completed a goal that I set for myself at the gym and it pumped me up for tomorrow. \n231                                                                                 I went to the gym with a friend of mine who I hadn't seen in a long time.\n555                                                                                                                         Getting to do yoga unexpectedly. \n...                                                                                                                                                       ...\n13447                                                                             I helped my friend Kyle create a workout regiment, he was really thankful. \n13456                                                                   I felt extra flexible today and was able to push into a deeper pose like I was Gumby.\n13734                                                                                                                                   I had a good workout.\n13843  I successfully rolled out the perpetual knots in my calves last night (after a week of work), and my posture and foot pain have improved drastically. \n14047                                                                                                                I exercise well today after a long time.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>122</th>\n      <td>Increased my max bench press this morning, heaviest I've ever lifted</td>\n    </tr>\n    <tr>\n      <th>147</th>\n      <td>I went to the gym. Hitting a new personal record on the bench press made me more confident in my abilities.</td>\n    </tr>\n    <tr>\n      <th>214</th>\n      <td>I completed a goal that I set for myself at the gym and it pumped me up for tomorrow.</td>\n    </tr>\n    <tr>\n      <th>231</th>\n      <td>I went to the gym with a friend of mine who I hadn't seen in a long time.</td>\n    </tr>\n    <tr>\n      <th>555</th>\n      <td>Getting to do yoga unexpectedly.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13447</th>\n      <td>I helped my friend Kyle create a workout regiment, he was really thankful.</td>\n    </tr>\n    <tr>\n      <th>13456</th>\n      <td>I felt extra flexible today and was able to push into a deeper pose like I was Gumby.</td>\n    </tr>\n    <tr>\n      <th>13734</th>\n      <td>I had a good workout.</td>\n    </tr>\n    <tr>\n      <th>13843</th>\n      <td>I successfully rolled out the perpetual knots in my calves last night (after a week of work), and my posture and foot pain have improved drastically.</td>\n    </tr>\n    <tr>\n      <th>14047</th>\n      <td>I exercise well today after a long time.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                154\nexercise                   147\naffection                    4\nbonding                      3\nenjoy_the_moment             1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>154</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>147</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>4</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>3</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 7 \tN Total: 222 (1.6%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                            Document\n11                                                                                                                      INDIA WON THE SERIES AGAINST AUSTRALIA MADE ME TO FEEL HAPPY\n52                                                                                                                                                     A good win for my sports team\n123                                                                                                                                          I watched a Blackhawks game last night.\n210                                                                                                                                            The team I coach won its match today.\n273    I made plans with my brother to get together this weekend to watch the Final Four.  My nephew could win $700, because he has Gonzaga beating Oregon in the championship game.\n...                                                                                                                                                                              ...\n13440                                                                                                                                                   I watched a basketball game.\n13445                                                                                                  I watched a double overtime game 7 in the Eastern Conference Hockey finals.  \n13531                                                                                                                                       The Seattle Mariners won last night 6-4.\n13690                                                                                                                   I was happy when the Mets came from behind to beat the Cubs.\n13715                                                                                                                         The red sox won and I watched it with friends at a bar\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>11</th>\n      <td>INDIA WON THE SERIES AGAINST AUSTRALIA MADE ME TO FEEL HAPPY</td>\n    </tr>\n    <tr>\n      <th>52</th>\n      <td>A good win for my sports team</td>\n    </tr>\n    <tr>\n      <th>123</th>\n      <td>I watched a Blackhawks game last night.</td>\n    </tr>\n    <tr>\n      <th>210</th>\n      <td>The team I coach won its match today.</td>\n    </tr>\n    <tr>\n      <th>273</th>\n      <td>I made plans with my brother to get together this weekend to watch the Final Four.  My nephew could win $700, because he has Gonzaga beating Oregon in the championship game.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13440</th>\n      <td>I watched a basketball game.</td>\n    </tr>\n    <tr>\n      <th>13445</th>\n      <td>I watched a double overtime game 7 in the Eastern Conference Hockey finals.</td>\n    </tr>\n    <tr>\n      <th>13531</th>\n      <td>The Seattle Mariners won last night 6-4.</td>\n    </tr>\n    <tr>\n      <th>13690</th>\n      <td>I was happy when the Mets came from behind to beat the Cubs.</td>\n    </tr>\n    <tr>\n      <th>13715</th>\n      <td>The red sox won and I watched it with friends at a bar</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nleisure                     65\nachievement                 51\nenjoy_the_moment            40\nbonding                     38\naffection                   20\nexercise                     8","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>leisure</th>\n      <td>65</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>51</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>40</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>38</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>20</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>8</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 8 \tN Total: 273 (1.9%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                     Document\n20     I was able to play my video game that I enjoy the most quite a bit and that made me relaxed and happy.\n104                                                               Did well in the videogame I was playing\\r\\n\n149                     I completed a modification for a video game that I was able to release to the public.\n167                                                I found something unexpected in a video game I am playing.\n442                                          Started a new game that I have been waiting a long time to play.\n...                                                                                                       ...\n13849                                                                Completing valkyr prime in a video game.\n13956   A small event that made me happy in the past 24 hours was the release of a Nintendo game called arms.\n14040                                              I made a great scorekeeping spreadsheet for a game I play.\n14093                                                                Purchasing and playing a new video game.\n14111                                         I was able to buy a really fun video game on sal at half price.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>20</th>\n      <td>I was able to play my video game that I enjoy the most quite a bit and that made me relaxed and happy.</td>\n    </tr>\n    <tr>\n      <th>104</th>\n      <td>Did well in the videogame I was playing\\r\\n</td>\n    </tr>\n    <tr>\n      <th>149</th>\n      <td>I completed a modification for a video game that I was able to release to the public.</td>\n    </tr>\n    <tr>\n      <th>167</th>\n      <td>I found something unexpected in a video game I am playing.</td>\n    </tr>\n    <tr>\n      <th>442</th>\n      <td>Started a new game that I have been waiting a long time to play.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13849</th>\n      <td>Completing valkyr prime in a video game.</td>\n    </tr>\n    <tr>\n      <th>13956</th>\n      <td>A small event that made me happy in the past 24 hours was the release of a Nintendo game called arms.</td>\n    </tr>\n    <tr>\n      <th>14040</th>\n      <td>I made a great scorekeeping spreadsheet for a game I play.</td>\n    </tr>\n    <tr>\n      <th>14093</th>\n      <td>Purchasing and playing a new video game.</td>\n    </tr>\n    <tr>\n      <th>14111</th>\n      <td>I was able to buy a really fun video game on sal at half price.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nleisure                    125\nachievement                 83\nbonding                     40\naffection                   13\nenjoy_the_moment            12","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>leisure</th>\n      <td>125</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>83</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>40</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>13</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>12</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 9 \tN Total: 224 (1.6%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                      Document\n335    I'm finally about to buy a car. My savings just got their few days ago. Now I'm looking. It's exciting.\n438                                                                                    i bought a new vehicle.\n726                                          I saw a really nice car on my way home, which is rare in my town.\n945                                                  I m very happy when i first time drive my own gifted car.\n1159                                                                       It was so amazing I get my new car.\n...                                                                                                        ...\n13227                                                                         Last day I got my new brand car.\n13301                                                       My friend took my car to get an oil change for me.\n13581                                                I test-drove a new vehicle that I ended up really liking.\n13899                                                                                       I BOUGHT A NEW CAR\n13981                                                             I went and bought some new tires for my car.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>335</th>\n      <td>I'm finally about to buy a car. My savings just got their few days ago. Now I'm looking. It's exciting.</td>\n    </tr>\n    <tr>\n      <th>438</th>\n      <td>i bought a new vehicle.</td>\n    </tr>\n    <tr>\n      <th>726</th>\n      <td>I saw a really nice car on my way home, which is rare in my town.</td>\n    </tr>\n    <tr>\n      <th>945</th>\n      <td>I m very happy when i first time drive my own gifted car.</td>\n    </tr>\n    <tr>\n      <th>1159</th>\n      <td>It was so amazing I get my new car.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13227</th>\n      <td>Last day I got my new brand car.</td>\n    </tr>\n    <tr>\n      <th>13301</th>\n      <td>My friend took my car to get an oil change for me.</td>\n    </tr>\n    <tr>\n      <th>13581</th>\n      <td>I test-drove a new vehicle that I ended up really liking.</td>\n    </tr>\n    <tr>\n      <th>13899</th>\n      <td>I BOUGHT A NEW CAR</td>\n    </tr>\n    <tr>\n      <th>13981</th>\n      <td>I went and bought some new tires for my car.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                179\naffection                   15\nenjoy_the_moment            14\nbonding                     14\nleisure                      2","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>179</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>15</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>14</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>14</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 10 \tN Total: 232 (1.6%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n17     After 3 long weeks of deep snow covered ground we finally had enough warm days in the Northeast for the snow to melt and I was able to start my day with a fantastic hike up the Appalachian trail t...\n81                                                                                                                                         I went for a walk early in the morning and air was clear and crisp.\n186                                                                                                      I stepped outside without a jacket today.  The sun is shining and it finally feels a bit like spring!\n189                                                                  I picked up running again now that the winter weather has finally died down. I feel much better physically and emotionally because of it.\n287                                                                                        The temperature today was 70 degrees and the sun was shining.  I don't like cold weather so this was heaven for me.\n...                                                                                                                                                                                                        ...\n13780                                                                                                                                  The wind blowing my hair while I stuck my head out of the car's window.\n13869                                                                                                                                                               I read outside for hours in the sunshine. \n13939                                                                                                                                                           There was a sudden shower of rain in our area.\n14027                                                                                                                                                                   It started to pour down rain and storm\n14041                                                                                                                                                 I felt more comfortable due to the weather cooling down.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>17</th>\n      <td>After 3 long weeks of deep snow covered ground we finally had enough warm days in the Northeast for the snow to melt and I was able to start my day with a fantastic hike up the Appalachian trail t...</td>\n    </tr>\n    <tr>\n      <th>81</th>\n      <td>I went for a walk early in the morning and air was clear and crisp.</td>\n    </tr>\n    <tr>\n      <th>186</th>\n      <td>I stepped outside without a jacket today.  The sun is shining and it finally feels a bit like spring!</td>\n    </tr>\n    <tr>\n      <th>189</th>\n      <td>I picked up running again now that the winter weather has finally died down. I feel much better physically and emotionally because of it.</td>\n    </tr>\n    <tr>\n      <th>287</th>\n      <td>The temperature today was 70 degrees and the sun was shining.  I don't like cold weather so this was heaven for me.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13780</th>\n      <td>The wind blowing my hair while I stuck my head out of the car's window.</td>\n    </tr>\n    <tr>\n      <th>13869</th>\n      <td>I read outside for hours in the sunshine.</td>\n    </tr>\n    <tr>\n      <th>13939</th>\n      <td>There was a sudden shower of rain in our area.</td>\n    </tr>\n    <tr>\n      <th>14027</th>\n      <td>It started to pour down rain and storm</td>\n    </tr>\n    <tr>\n      <th>14041</th>\n      <td>I felt more comfortable due to the weather cooling down.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nnature                     168\nenjoy_the_moment            19\nexercise                    13\nachievement                 10\naffection                    9\nleisure                      8\nbonding                      5","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>nature</th>\n      <td>168</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>19</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>13</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>10</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>9</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>8</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>5</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 11 \tN Total: 205 (1.5%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                            Document\n1                                                                                                            I meditated last night.\n120                                    I got a full night of sleep. That does not often happen with a 3 month-old in the house. \\r\\n\n169                                                                                                I got to take a nap on the couch.\n213                                                                                             I had an amazing dream this morning.\n480                                                                     When I jumped in bed at night after a fairly exhaustive day.\n...                                                                                                                              ...\n13846                                  I took a really nice nap while the air conditioner was on and woke up feeling very refreshed.\n13852                                                                                      Getting a good night sleep made me happy.\n13970                                                                                        I was able to get a great nights sleep.\n14063                                                                           I WAS SLEEPING IN LONG TIME. SO I FELT VERY RELAXED.\n14122  I woke up in the middle of the night and realizing that I still had another four hours of sleep left before having to get up.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1</th>\n      <td>I meditated last night.</td>\n    </tr>\n    <tr>\n      <th>120</th>\n      <td>I got a full night of sleep. That does not often happen with a 3 month-old in the house. \\r\\n</td>\n    </tr>\n    <tr>\n      <th>169</th>\n      <td>I got to take a nap on the couch.</td>\n    </tr>\n    <tr>\n      <th>213</th>\n      <td>I had an amazing dream this morning.</td>\n    </tr>\n    <tr>\n      <th>480</th>\n      <td>When I jumped in bed at night after a fairly exhaustive day.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13846</th>\n      <td>I took a really nice nap while the air conditioner was on and woke up feeling very refreshed.</td>\n    </tr>\n    <tr>\n      <th>13852</th>\n      <td>Getting a good night sleep made me happy.</td>\n    </tr>\n    <tr>\n      <th>13970</th>\n      <td>I was able to get a great nights sleep.</td>\n    </tr>\n    <tr>\n      <th>14063</th>\n      <td>I WAS SLEEPING IN LONG TIME. SO I FELT VERY RELAXED.</td>\n    </tr>\n    <tr>\n      <th>14122</th>\n      <td>I woke up in the middle of the night and realizing that I still had another four hours of sleep left before having to get up.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nenjoy_the_moment            85\nleisure                     53\nachievement                 39\naffection                   22\nexercise                     5\nnature                       1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>85</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>53</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>39</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>22</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>5</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 12 \tN Total: 427 (3.0%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                      Document\n14     A made quite the progress on an old rusty workbench I am restoring at the moment. With the new paint it looks like new.\n67                          I managed to throw out a few items that were cluttering a room. I enjoyed the space I have after. \n91                                                                                         I finally cleaned out my basement. \n157                    A friend came over to help install an appliance.  I was very happy I didn't have to pay a professional.\n174                                           I completed upgrading the switches in our house so everything looks much better.\n...                                                                                                                        ...\n13562                                                                                                  Cleaned the garage out.\n13685                                                                                    I bought a new appliance for kitchen.\n13700                                                                             Got halfway done painting the guest bedroom.\n13944                                                                          I found a new, simple way to clean some things.\n14116                                         I made our new house that we just bought look better by getting things put away.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>14</th>\n      <td>A made quite the progress on an old rusty workbench I am restoring at the moment. With the new paint it looks like new.</td>\n    </tr>\n    <tr>\n      <th>67</th>\n      <td>I managed to throw out a few items that were cluttering a room. I enjoyed the space I have after.</td>\n    </tr>\n    <tr>\n      <th>91</th>\n      <td>I finally cleaned out my basement.</td>\n    </tr>\n    <tr>\n      <th>157</th>\n      <td>A friend came over to help install an appliance.  I was very happy I didn't have to pay a professional.</td>\n    </tr>\n    <tr>\n      <th>174</th>\n      <td>I completed upgrading the switches in our house so everything looks much better.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13562</th>\n      <td>Cleaned the garage out.</td>\n    </tr>\n    <tr>\n      <th>13685</th>\n      <td>I bought a new appliance for kitchen.</td>\n    </tr>\n    <tr>\n      <th>13700</th>\n      <td>Got halfway done painting the guest bedroom.</td>\n    </tr>\n    <tr>\n      <th>13944</th>\n      <td>I found a new, simple way to clean some things.</td>\n    </tr>\n    <tr>\n      <th>14116</th>\n      <td>I made our new house that we just bought look better by getting things put away.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                331\nenjoy_the_moment            51\naffection                   23\nleisure                     11\nbonding                     10\nnature                       1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>331</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>51</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>23</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>11</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>10</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 13 \tN Total: 148 (1.0%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                      Document\n225                                                                                                           I was able to turn in the forbearance form for my student loans.\n446                                                                                                                                 That I finally paid off my hospital bills!\n583                                                                                                                                                      I got my bills paid. \n648                                                                         I checked my bank account after I was paid via direct deposit, and made more than I thought I did.\n713                                                                           I finally completed the process of divesting from a big bank with unethical investing practices.\n...                                                                                                                                                                        ...\n13587                                                                     I received a refund check from an account that I had at least 10 years ago and forgot that I had it.\n13619                                               I got a refund of $99.00 from an online newspaper which, three days ago, removed the same amount from my checking account.\n13649                                                                                                                           My electric bill was $20 less than I expected.\n13796                                                                                                                                I transferred money into my bank account.\n14090  I was supposed to send a payment to someone who made the payment on my behalf. It is mid-month and was difficult, but we managed to send the funds. This made me happy.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>225</th>\n      <td>I was able to turn in the forbearance form for my student loans.</td>\n    </tr>\n    <tr>\n      <th>446</th>\n      <td>That I finally paid off my hospital bills!</td>\n    </tr>\n    <tr>\n      <th>583</th>\n      <td>I got my bills paid.</td>\n    </tr>\n    <tr>\n      <th>648</th>\n      <td>I checked my bank account after I was paid via direct deposit, and made more than I thought I did.</td>\n    </tr>\n    <tr>\n      <th>713</th>\n      <td>I finally completed the process of divesting from a big bank with unethical investing practices.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13587</th>\n      <td>I received a refund check from an account that I had at least 10 years ago and forgot that I had it.</td>\n    </tr>\n    <tr>\n      <th>13619</th>\n      <td>I got a refund of $99.00 from an online newspaper which, three days ago, removed the same amount from my checking account.</td>\n    </tr>\n    <tr>\n      <th>13649</th>\n      <td>My electric bill was $20 less than I expected.</td>\n    </tr>\n    <tr>\n      <th>13796</th>\n      <td>I transferred money into my bank account.</td>\n    </tr>\n    <tr>\n      <th>14090</th>\n      <td>I was supposed to send a payment to someone who made the payment on my behalf. It is mid-month and was difficult, but we managed to send the funds. This made me happy.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                130\nenjoy_the_moment             9\naffection                    6\nbonding                      3","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>130</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>9</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>6</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>3</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 14 \tN Total: 233 (1.6%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n106                                                                                                                I got done with a test that I had been nervous about taking. I think I did pretty good too.\n336                                                                                                                   I bought several books for cheap that would allow me to study well for an upcoming test.\n347                                             My happiness had its complement when I was called to congratulate myself for my graduation of an intensive course of programming. I felt expert in technology.\n353                          I got the invigilation duty for class X board exams this year.  This is the first time get the duty after 19 years of work.  This has given me a tremendous joy and satisfaction.\n421                                                                  I found out that I passed my qualifying exam at my graduate school, which was a huge hurdle I had to overcome and was very nervous about.\n...                                                                                                                                                                                                        ...\n13207                                                                                                                                                                        when my test result was published\n13244  I was taking a quiz for my algebra course online, and I scored an 83%. I've never been great at math and my extra reading and practice is starting to pay off. I couldn't stop smiling, and I had to...\n13295                                                                                                                                                 I was happy when I found out I got 88% on my quiz today.\n13540                                                                                                                                 I just completed my boating licence exam so now I can ride wave runners!\n13595                                                                                                                                                                                Passed my CPA Examination\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>106</th>\n      <td>I got done with a test that I had been nervous about taking. I think I did pretty good too.</td>\n    </tr>\n    <tr>\n      <th>336</th>\n      <td>I bought several books for cheap that would allow me to study well for an upcoming test.</td>\n    </tr>\n    <tr>\n      <th>347</th>\n      <td>My happiness had its complement when I was called to congratulate myself for my graduation of an intensive course of programming. I felt expert in technology.</td>\n    </tr>\n    <tr>\n      <th>353</th>\n      <td>I got the invigilation duty for class X board exams this year.  This is the first time get the duty after 19 years of work.  This has given me a tremendous joy and satisfaction.</td>\n    </tr>\n    <tr>\n      <th>421</th>\n      <td>I found out that I passed my qualifying exam at my graduate school, which was a huge hurdle I had to overcome and was very nervous about.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13207</th>\n      <td>when my test result was published</td>\n    </tr>\n    <tr>\n      <th>13244</th>\n      <td>I was taking a quiz for my algebra course online, and I scored an 83%. I've never been great at math and my extra reading and practice is starting to pay off. I couldn't stop smiling, and I had to...</td>\n    </tr>\n    <tr>\n      <th>13295</th>\n      <td>I was happy when I found out I got 88% on my quiz today.</td>\n    </tr>\n    <tr>\n      <th>13540</th>\n      <td>I just completed my boating licence exam so now I can ride wave runners!</td>\n    </tr>\n    <tr>\n      <th>13595</th>\n      <td>Passed my CPA Examination</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                221\nenjoy_the_moment             6\naffection                    2\nbonding                      2\nexercise                     1\nleisure                      1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>221</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>6</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 15 \tN Total: 1019 (7.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n92                                                                                                                                                                   My new boss appreciated the work i put in\n124                                                                                                                                                                        complete my pending job task today.\n150                                                                                                                                          I finally figured out how to properly install a program at work. \n159                                                                                                                                                        I finished the work project I was working on today.\n170                                                                                                        I participated in a group task and felt valued in it, as if my skills were needed and appreciated. \n...                                                                                                                                                                                                        ...\n13827                                                                                                                                                I completed a large assignment at work ahead of schedule.\n13844                                                                                                                                                       I made a somewhat significant milestone in my job.\n13859                                                                                                                    I finished a project at work that I had been trying to get done for a several months.\n13889                                                                                                                                                                            I finished a project at work.\n14051  Every month, I host breakfast with 10 employees from different levels and functions in our company a from creative to business development to data science to marketing; it's always a unique and fu...\n\n[194 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>92</th>\n      <td>My new boss appreciated the work i put in</td>\n    </tr>\n    <tr>\n      <th>124</th>\n      <td>complete my pending job task today.</td>\n    </tr>\n    <tr>\n      <th>150</th>\n      <td>I finally figured out how to properly install a program at work.</td>\n    </tr>\n    <tr>\n      <th>159</th>\n      <td>I finished the work project I was working on today.</td>\n    </tr>\n    <tr>\n      <th>170</th>\n      <td>I participated in a group task and felt valued in it, as if my skills were needed and appreciated.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13827</th>\n      <td>I completed a large assignment at work ahead of schedule.</td>\n    </tr>\n    <tr>\n      <th>13844</th>\n      <td>I made a somewhat significant milestone in my job.</td>\n    </tr>\n    <tr>\n      <th>13859</th>\n      <td>I finished a project at work that I had been trying to get done for a several months.</td>\n    </tr>\n    <tr>\n      <th>13889</th>\n      <td>I finished a project at work.</td>\n    </tr>\n    <tr>\n      <th>14051</th>\n      <td>Every month, I host breakfast with 10 employees from different levels and functions in our company a from creative to business development to data science to marketing; it's always a unique and fu...</td>\n    </tr>\n  </tbody>\n</table>\n<p>194 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nachievement                878\nenjoy_the_moment            64\nbonding                     38\naffection                   18\nleisure                     17\nnature                       2\nexercise                     2","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>achievement</th>\n      <td>878</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>64</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>38</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>18</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>17</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>exercise</th>\n      <td>2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 16 \tN Total: 686 (4.9%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n323                                                                                                                                                My youngest son began to stand up holding on to an object. \n338                                                                          I was treating my sons family to Sunday lunch and I caught my very intelligent 16 year old grandson off guard and made him laugh.\n362                                                                                                                            It made me happy when my infant son used baby sign language for the first time.\n386                                                                                                                              My daughter started walking, and has gotten much better in the past few days.\n524                                                                                                                                                                                       My baby smiled at me\n...                                                                                                                                                                                                        ...\n13692  I felt happy when I took my daughter to her drum lesson and saw her smiling when I picked her up. She told me she had a great lesson and learned something new and she thanked me for taking her to ...\n13842                                                                                       My younger son comforted my older son after he got hurt. He was demonstrating extreme empathy at only 6 years old.\n13965                                                                                                                                                                    MY son started taking his first steps\n13986                                 Two year old Kaarmikha, my daughter started talking some new words in her own language. Kissing me often and demanding her needs. Motherhood is the best thing in world.\n14030                                                          I have a daughter and she is 3 years old. Today she had made some remarks which made feel very happy. Her talks really gives me a sense of joy.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>323</th>\n      <td>My youngest son began to stand up holding on to an object.</td>\n    </tr>\n    <tr>\n      <th>338</th>\n      <td>I was treating my sons family to Sunday lunch and I caught my very intelligent 16 year old grandson off guard and made him laugh.</td>\n    </tr>\n    <tr>\n      <th>362</th>\n      <td>It made me happy when my infant son used baby sign language for the first time.</td>\n    </tr>\n    <tr>\n      <th>386</th>\n      <td>My daughter started walking, and has gotten much better in the past few days.</td>\n    </tr>\n    <tr>\n      <th>524</th>\n      <td>My baby smiled at me</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13692</th>\n      <td>I felt happy when I took my daughter to her drum lesson and saw her smiling when I picked her up. She told me she had a great lesson and learned something new and she thanked me for taking her to ...</td>\n    </tr>\n    <tr>\n      <th>13842</th>\n      <td>My younger son comforted my older son after he got hurt. He was demonstrating extreme empathy at only 6 years old.</td>\n    </tr>\n    <tr>\n      <th>13965</th>\n      <td>MY son started taking his first steps</td>\n    </tr>\n    <tr>\n      <th>13986</th>\n      <td>Two year old Kaarmikha, my daughter started talking some new words in her own language. Kissing me often and demanding her needs. Motherhood is the best thing in world.</td>\n    </tr>\n    <tr>\n      <th>14030</th>\n      <td>I have a daughter and she is 3 years old. Today she had made some remarks which made feel very happy. Her talks really gives me a sense of joy.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  603\nachievement                 60\nbonding                     13\nenjoy_the_moment             6\nleisure                      3\nnature                       1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>603</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>60</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>13</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>6</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>3</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 17 \tN Total: 717 (5.1%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                              Document\n10                                                                               i finally learned to cook chicken\\r\\n\n46                                                                      My job provided lunch and it was really good. \n71                                                                       I satisfied my many food cravings for dinner.\n113                                                                        I got Chinese food for dinner, it was good.\n127                    I had cream of chicken soup for dinner. It made me feel a lot better because I have been sick. \n...                                                                                                                ...\n13656                                      I had a great meal of grilled chicken, soup, and salad that was delicious. \n13741                                                                                              I ate a nice lunch.\n13979                                                                                                 I eat good food.\n14021  I nan some chicken skewers with a delicious teriyaki sauce. I am always in my happy place when I man the grill.\n14108                                                                  I ate some delicious eggrolls that my mom made.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>10</th>\n      <td>i finally learned to cook chicken\\r\\n</td>\n    </tr>\n    <tr>\n      <th>46</th>\n      <td>My job provided lunch and it was really good.</td>\n    </tr>\n    <tr>\n      <th>71</th>\n      <td>I satisfied my many food cravings for dinner.</td>\n    </tr>\n    <tr>\n      <th>113</th>\n      <td>I got Chinese food for dinner, it was good.</td>\n    </tr>\n    <tr>\n      <th>127</th>\n      <td>I had cream of chicken soup for dinner. It made me feel a lot better because I have been sick.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13656</th>\n      <td>I had a great meal of grilled chicken, soup, and salad that was delicious.</td>\n    </tr>\n    <tr>\n      <th>13741</th>\n      <td>I ate a nice lunch.</td>\n    </tr>\n    <tr>\n      <th>13979</th>\n      <td>I eat good food.</td>\n    </tr>\n    <tr>\n      <th>14021</th>\n      <td>I nan some chicken skewers with a delicious teriyaki sauce. I am always in my happy place when I man the grill.</td>\n    </tr>\n    <tr>\n      <th>14108</th>\n      <td>I ate some delicious eggrolls that my mom made.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nenjoy_the_moment           299\nachievement                157\naffection                  119\nleisure                    101\nbonding                     39\nnature                       2","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>299</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>157</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>119</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>101</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>39</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 18 \tN Total: 168 (1.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n48                                                                                                                                           I made vacation plans with my daughter today for Florida in July.\n233    I am planning another trip upcoming in August. This one is a week-long trip in Colorado. Planning makes me so excited. I love seeing everything coming together and falling into place. I will proba...\n340                                                                                                                                                             I traveled to a state I had never been before.\n383                                                                                                                       I found out that I had enough money to take a vacation this summer to the Caribbean!\n388                                                                                                                                                      I successfully learned how to hover in a helicopter. \n...                                                                                                                                                                                                        ...\n12992                                                                                                            yesterday night i came to trichy to erode is very nice travel and thankful to god nice moment\n13168                                                                                                                                         I went my friend's with beach and  it is very entertaining  one.\n13532                                                                                                                               We went to a meeting to finalize plans for my daughter to travel to Spain.\n13806                                                                                                                                                    I booked a weekend trip with my friends to the beach.\n13907                                                         LAST SUMMER HOLIDAYS WAS AN UNFORGETTABLE EXPERIENCE.  NT FRIEND ALWAYS ENJOYED MANY places.  WE ENJOYED OUR EXCURSION WITH ENJOY AND BEAUTIFUL.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>48</th>\n      <td>I made vacation plans with my daughter today for Florida in July.</td>\n    </tr>\n    <tr>\n      <th>233</th>\n      <td>I am planning another trip upcoming in August. This one is a week-long trip in Colorado. Planning makes me so excited. I love seeing everything coming together and falling into place. I will proba...</td>\n    </tr>\n    <tr>\n      <th>340</th>\n      <td>I traveled to a state I had never been before.</td>\n    </tr>\n    <tr>\n      <th>383</th>\n      <td>I found out that I had enough money to take a vacation this summer to the Caribbean!</td>\n    </tr>\n    <tr>\n      <th>388</th>\n      <td>I successfully learned how to hover in a helicopter.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>12992</th>\n      <td>yesterday night i came to trichy to erode is very nice travel and thankful to god nice moment</td>\n    </tr>\n    <tr>\n      <th>13168</th>\n      <td>I went my friend's with beach and  it is very entertaining  one.</td>\n    </tr>\n    <tr>\n      <th>13532</th>\n      <td>We went to a meeting to finalize plans for my daughter to travel to Spain.</td>\n    </tr>\n    <tr>\n      <th>13806</th>\n      <td>I booked a weekend trip with my friends to the beach.</td>\n    </tr>\n    <tr>\n      <th>13907</th>\n      <td>LAST SUMMER HOLIDAYS WAS AN UNFORGETTABLE EXPERIENCE.  NT FRIEND ALWAYS ENJOYED MANY places.  WE ENJOYED OUR EXCURSION WITH ENJOY AND BEAUTIFUL.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                   57\nachievement                 32\nbonding                     27\nleisure                     26\nenjoy_the_moment            23\nnature                       3","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>57</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>32</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>27</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>26</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>23</td>\n    </tr>\n    <tr>\n      <th>nature</th>\n      <td>3</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 19 \tN Total: 173 (1.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                           Document\n2                                                                      My grandmother start to walk from the bed after a long time.\n66                                                     After a long time, today my cousin came home surprisingly that made my day !\n314                                                                       Going home to see my family after being gone for college.\n706    My parents trying to be super supportive even if their ideas weren't always helpful because I could see how much they cared.\n842                                                   YESTERDAY I WAS GONE FOR MY RELATIVE FUNCTION, BECAUSE MEET MY ALL RELATIVES.\n...                                                                                                                             ...\n13216                                                                      I got to see my grandma after 5 months of her vacation. \n13359                                                                I was happy when I talked to my parents earlier this morning. \n13395                                                                   Came home to find my parents at my house for a quick visit.\n13798                                   Yesterday I found out my sister is coming to visit me this weekend, and that made me happy.\n14123                                                  Yesterday my relations came to my house. That time am very happy to saw them\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2</th>\n      <td>My grandmother start to walk from the bed after a long time.</td>\n    </tr>\n    <tr>\n      <th>66</th>\n      <td>After a long time, today my cousin came home surprisingly that made my day !</td>\n    </tr>\n    <tr>\n      <th>314</th>\n      <td>Going home to see my family after being gone for college.</td>\n    </tr>\n    <tr>\n      <th>706</th>\n      <td>My parents trying to be super supportive even if their ideas weren't always helpful because I could see how much they cared.</td>\n    </tr>\n    <tr>\n      <th>842</th>\n      <td>YESTERDAY I WAS GONE FOR MY RELATIVE FUNCTION, BECAUSE MEET MY ALL RELATIVES.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13216</th>\n      <td>I got to see my grandma after 5 months of her vacation.</td>\n    </tr>\n    <tr>\n      <th>13359</th>\n      <td>I was happy when I talked to my parents earlier this morning.</td>\n    </tr>\n    <tr>\n      <th>13395</th>\n      <td>Came home to find my parents at my house for a quick visit.</td>\n    </tr>\n    <tr>\n      <th>13798</th>\n      <td>Yesterday I found out my sister is coming to visit me this weekend, and that made me happy.</td>\n    </tr>\n    <tr>\n      <th>14123</th>\n      <td>Yesterday my relations came to my house. That time am very happy to saw them</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  168\nachievement                  2\nenjoy_the_moment             2\nbonding                      1","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>168</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 20 \tN Total: 492 (3.5%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n79                                                                                                                                       I talked to my girlfriend recently and she told me that she loved me.\n129                                                                                                                           Picking up my girlfriend from the airport since I hadn't seen her for two weeks.\n212                                                                                                                                                    I saw my boyfriend again after being apart for a month.\n278                                                                                                                            I got my bff to stop yelling at me all the time because her boyfriend is a dick\n502                                                                                                                                                My lover presented one precious gift to me made very happy.\n...                                                                                                                                                                                                        ...\n13671                                          My boyfriend unexpectedly surprised me by visiting at my parents' house where I am dogsitting. I wasn't expecting to see him last night, but he came by anyway!\n13699                                                                                                                                                                         Met my girlfriend after 15 days.\n13874                                                                                                                                                            I got to spend the evening with my boyfriend.\n13893  I came to a ton of conclusions which all amounted to if he loved me, he would show me, consistently, the way I wanted to be shown love. Simple things like love letters, flowers, romantic date nigh...\n14124                                                                                                                                                                  Cuddling with my girlfriend last night.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>79</th>\n      <td>I talked to my girlfriend recently and she told me that she loved me.</td>\n    </tr>\n    <tr>\n      <th>129</th>\n      <td>Picking up my girlfriend from the airport since I hadn't seen her for two weeks.</td>\n    </tr>\n    <tr>\n      <th>212</th>\n      <td>I saw my boyfriend again after being apart for a month.</td>\n    </tr>\n    <tr>\n      <th>278</th>\n      <td>I got my bff to stop yelling at me all the time because her boyfriend is a dick</td>\n    </tr>\n    <tr>\n      <th>502</th>\n      <td>My lover presented one precious gift to me made very happy.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13671</th>\n      <td>My boyfriend unexpectedly surprised me by visiting at my parents' house where I am dogsitting. I wasn't expecting to see him last night, but he came by anyway!</td>\n    </tr>\n    <tr>\n      <th>13699</th>\n      <td>Met my girlfriend after 15 days.</td>\n    </tr>\n    <tr>\n      <th>13874</th>\n      <td>I got to spend the evening with my boyfriend.</td>\n    </tr>\n    <tr>\n      <th>13893</th>\n      <td>I came to a ton of conclusions which all amounted to if he loved me, he would show me, consistently, the way I wanted to be shown love. Simple things like love letters, flowers, romantic date nigh...</td>\n    </tr>\n    <tr>\n      <th>14124</th>\n      <td>Cuddling with my girlfriend last night.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  457\nbonding                     21\nachievement                  8\nenjoy_the_moment             4\nleisure                      2","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>457</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>21</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>8</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>4</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 21 \tN Total: 163 (1.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                                                                      Document\n26                                                                                                                                                         I finally decorated my son's room for his birthday.\n57                                                                                                                               I got to see a very good old friend that came over at my daughter's birthday.\n117                                                                                                                                I helped plan a surprise party for my Mom's 70th birthday party coming up. \n328    Yesterday was my birthday and usually I am not honored or get more than a general happy birthday from my family and friends.  But my boyfriend stayed up until one minute after midnight the night b...\n350                                                                                                                                             My husband and me fighting we are not celebrating my birthday \n...                                                                                                                                                                                                        ...\n13483                                                                                                                       My middle son's birthday was this week and he was so anxious and excited about it.\n13493                                                                                                                                                     It was my son's birthday and we had a party for him.\n13990                                                                                                                                                                when my nephew got celebrate his birthday\n14033                                                                                                        I have celebrated my mom's birthday today with my family and friends and that made me very happy.\n14064                                                                                                                                                                            I CELEBRATED MY SON BIRTHDAY.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>26</th>\n      <td>I finally decorated my son's room for his birthday.</td>\n    </tr>\n    <tr>\n      <th>57</th>\n      <td>I got to see a very good old friend that came over at my daughter's birthday.</td>\n    </tr>\n    <tr>\n      <th>117</th>\n      <td>I helped plan a surprise party for my Mom's 70th birthday party coming up.</td>\n    </tr>\n    <tr>\n      <th>328</th>\n      <td>Yesterday was my birthday and usually I am not honored or get more than a general happy birthday from my family and friends.  But my boyfriend stayed up until one minute after midnight the night b...</td>\n    </tr>\n    <tr>\n      <th>350</th>\n      <td>My husband and me fighting we are not celebrating my birthday</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13483</th>\n      <td>My middle son's birthday was this week and he was so anxious and excited about it.</td>\n    </tr>\n    <tr>\n      <th>13493</th>\n      <td>It was my son's birthday and we had a party for him.</td>\n    </tr>\n    <tr>\n      <th>13990</th>\n      <td>when my nephew got celebrate his birthday</td>\n    </tr>\n    <tr>\n      <th>14033</th>\n      <td>I have celebrated my mom's birthday today with my family and friends and that made me very happy.</td>\n    </tr>\n    <tr>\n      <th>14064</th>\n      <td>I CELEBRATED MY SON BIRTHDAY.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\naffection                  105\nbonding                     36\nenjoy_the_moment            18\nachievement                  2\nleisure                      2","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>affection</th>\n      <td>105</td>\n    </tr>\n    <tr>\n      <th>bonding</th>\n      <td>36</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>18</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>2</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\nCluster: 22 \tN Total: 447 (3.2%)\nCluster Exemplars: \n"},{"output_type":"display_data","data":{"text/plain":"                                                                                                                                                   Document\n64                                                 i met my old friends after 5 years.we all had a talk for half a day.we played football like my old days.\n93                                                                         I saw some friends that I had not seen in a while and it was good to reconnect. \n300                                                                                                                  I spent some time with a close friend.\n354                                                                                                         I met with my school friends and enjoyed a lot.\n355                                                                I got to see one of my friends that I havent seen in a long time and we got to catch up.\n...                                                                                                                                                     ...\n13718                                                                                                                    Today I meet my childhood friends.\n13745                                                                                                                                  Visit my best friend\n13771                                                                                               A friend called me and we caught up on a lot of things.\n13839  I got to see my best friend for the first time in over a month. I've really missed her - like missing a part of myself (my better half practically.)\n13881                                                                                                                         My friend came over to visit.\n\n[141 rows x 1 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Document</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>64</th>\n      <td>i met my old friends after 5 years.we all had a talk for half a day.we played football like my old days.</td>\n    </tr>\n    <tr>\n      <th>93</th>\n      <td>I saw some friends that I had not seen in a while and it was good to reconnect.</td>\n    </tr>\n    <tr>\n      <th>300</th>\n      <td>I spent some time with a close friend.</td>\n    </tr>\n    <tr>\n      <th>354</th>\n      <td>I met with my school friends and enjoyed a lot.</td>\n    </tr>\n    <tr>\n      <th>355</th>\n      <td>I got to see one of my friends that I havent seen in a long time and we got to catch up.</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>13718</th>\n      <td>Today I meet my childhood friends.</td>\n    </tr>\n    <tr>\n      <th>13745</th>\n      <td>Visit my best friend</td>\n    </tr>\n    <tr>\n      <th>13771</th>\n      <td>A friend called me and we caught up on a lot of things.</td>\n    </tr>\n    <tr>\n      <th>13839</th>\n      <td>I got to see my best friend for the first time in over a month. I've really missed her - like missing a part of myself (my better half practically.)</td>\n    </tr>\n    <tr>\n      <th>13881</th>\n      <td>My friend came over to visit.</td>\n    </tr>\n  </tbody>\n</table>\n<p>141 rows × 1 columns</p>\n</div>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"                  Ground Truth\nbonding                    415\naffection                   12\nachievement                  9\nenjoy_the_moment             7\nleisure                      4","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Ground Truth</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>bonding</th>\n      <td>415</td>\n    </tr>\n    <tr>\n      <th>affection</th>\n      <td>12</td>\n    </tr>\n    <tr>\n      <th>achievement</th>\n      <td>9</td>\n    </tr>\n    <tr>\n      <th>enjoy_the_moment</th>\n      <td>7</td>\n    </tr>\n    <tr>\n      <th>leisure</th>\n      <td>4</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"-------------------------\n"}],"source":["for key, item in grouped_df:\n","    n =  len(grouped_df.get_group(key))\n","    if key == -1:\n","        print(f\"No Cluster (outlier) \\tN Total: {n} ({(n/len(embeddings_df)) * 100:.1f}%)\")\n","        print(\"-----\" * 5)\n","        continue\n","    print(f\"Cluster: {key} \\tN Total: {n} ({(n/len(embeddings_df)) * 100:.1f}%)\")\n","    print(\"Cluster Exemplars: \")\n","    display(grouped_df.get_group(key).query(\"Exemplar\")[\"Document\"].to_frame())\n","    display(grouped_df.get_group(key)[\"Ground Truth\"].value_counts().to_frame())\n","    print(\"-----\" * 5)"]},{"cell_type":"code","execution_count":null,"metadata":{},"outputs":[],"source":[]}],"metadata":{"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.7.5-final"},"orig_nbformat":2,"kernelspec":{"name":"python37564bitanalysisconda0307f525ce98408fa0d9c21d055c6dba","display_name":"Python 3.7.5 64-bit ('analysis': conda)"}},"nbformat":4,"nbformat_minor":2}


--------------------------------------------------------------------------------
/pipfreeze.sh:
--------------------------------------------------------------------------------
1 | pip freeze > binder/requirements.txt
2 | grep -v "en-core-web" binder/requirements.txt > requirements; mv requirements binder/requirements.txt
3 | grep -v "certifi" binder/requirements.txt > requirements; mv requirements binder/requirements.txt
4 | 


--------------------------------------------------------------------------------
/readme.md:
--------------------------------------------------------------------------------
1 | # Notebooks
2 | 
3 | [![Binder](https://mybinder.org/badge.svg)](https://mybinder.org/v2/gh/pmbaumgartner/binder-notebooks/master)
4 | 
5 | These are notebooks for some of the longer workflows I've developed. You can launch them with [Binder](https://mybinder.org/v2/gh/pmbaumgartner/binder-notebooks/master)!
6 | 
7 | If you want to replicate these notebooks locally, there is a `requirements.txt` file is in the `binder` folder. Use this to install the required packages into a virtual environment.
8 | 
9 | Note that some notebooks are nondeterministic because I failed to set a random seed at some points. Thus, the exact results you get might differ from what you see in the notebooks if you run them again.


--------------------------------------------------------------------------------
/roc_colored.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/pmbaumgartner/binder-notebooks/cfee899187c9ba5e7a11a5b26fb37aa78cbc1201/roc_colored.png


--------------------------------------------------------------------------------