├── README_files
    └── figure-gfm
    │   ├── main_plot-1.png
    │   └── pride_plot-1.png
├── README.md
├── scripts
    └── the_emotions_of_jane_austen.R
└── README.Rmd


/README_files/figure-gfm/main_plot-1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/filmicaesthetic/JaneAustenStreamgraphs/HEAD/README_files/figure-gfm/main_plot-1.png


--------------------------------------------------------------------------------
/README_files/figure-gfm/pride_plot-1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/filmicaesthetic/JaneAustenStreamgraphs/HEAD/README_files/figure-gfm/pride_plot-1.png


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | The Emotions of Jane Austen
 2 | ================
 3 | 
 4 | This personal project explores the literature of Jane Austen novels,
 5 | provided by Julia Silge’s
 6 | <a href="https://github.com/juliasilge/janeaustenr">janeaustenr</a>
 7 | package in R, using the
 8 | <a href="https://saifmohammad.com/WebPages/NRC-Emotion-Lexicon.htm">NRC
 9 | Word-Emotion Association Lexicon</a> by Saif Mohammad. The lexicon is
10 | used to identify words associated with a range of 8 emotions: joy,
11 | trust, surprise, anticipation, sadness, fear, anger, and disgust.
12 | 
13 | The project aims to capture a visual interpretation of the emotional
14 | journey in each novel with a single visualisation.
15 | 
16 | ![](README_files/figure-gfm/main_plot-1.png)<!-- -->
17 | 
18 | ## Files
19 | 
20 | ‘the_emotions_of_jane_austen.R’ :
21 | <a href="https://github.com/filmicaesthetic/JaneAustenStreamgraphs/blob/main/scripts/the_emotions_of_jane_austen.R">view
22 | code</a>
23 | 
24 | ## Process
25 | 
26 | -   Load the NRC Word-Emotion Association Lexicon and Jane Austen novels
27 |     datasets and prepare for further analysis.
28 | -   Count occurrences of words with an emotional association for each
29 |     line of each book.
30 | -   Group lines approximately into pages (30 lines) to help reduce noise
31 |     in the final visualisation.
32 | -   Identify the two emotions with strongest association for each group.
33 | -   Plot the results for each novel in a streamgraph.
34 | -   Create grid of all plots and bookshelf-style legend.
35 | 


--------------------------------------------------------------------------------
/scripts/the_emotions_of_jane_austen.R:
--------------------------------------------------------------------------------
  1 | ###########
  2 | #
  3 | # Analysis of the emotion of language used in Jane Austen novels,
  4 | # using the NRC Emotion Lexicon.
  5 | #
  6 | ###########
  7 | 
  8 | # load packages
  9 | library(plyr)
 10 | library(stringr)
 11 | library(data.table)
 12 | library(ggplot2)
 13 | library(ggstream)
 14 | library(tidytext)
 15 | library(janeaustenr)
 16 | library(extrafont)
 17 | library(showtext)
 18 | library(dplyr)
 19 | library(gridExtra)
 20 | library(grid)
 21 | library(cowplot)
 22 | 
 23 | # format the data from {janeaustenr} for analysis
 24 | tidy_books <- austen_books() %>%
 25 |   group_by(book) %>%
 26 |   mutate(
 27 |     linenumber = row_number(),
 28 |     chapter = cumsum(str_detect(text, 
 29 |                                 regex("^chapter [\\divxlc]", 
 30 |                                       ignore_case = TRUE)))) %>%
 31 |   ungroup() %>%
 32 |   unnest_tokens(word, text)
 33 | 
 34 | # import NRC Word-Emotion Association Sentiment Lexicon
 35 | nrc_all <- get_sentiments("nrc")
 36 | 
 37 | # list of unique emotions
 38 | sentiments <- unique(nrc_all$sentiment)
 39 | 
 40 | # identify words that appear in the emotion lexicon
 41 | all_books <- tidy_books %>% 
 42 |   filter() %>%
 43 |   mutate(joy = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "joy"]),
 44 |          trust = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "trust"]),
 45 |          surprise = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "surprise"]),
 46 |          anticipation = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "anticipation"]),
 47 |          sadness = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "sadness"]),
 48 |          fear = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "fear"]),
 49 |          anger = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "anger"]),
 50 |          disgust = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "disgust"]))
 51 | 
 52 | # reshape the data frame
 53 | all_books_long <- melt(setDT(all_books), id.vars = c("book","linenumber","chapter","word"), variable.name = "emotion")
 54 | # round lines to nearest 30 (approximately 1 page) to help remove noise
 55 | all_books_long_10 <- all_books_long %>% mutate(linenumber = round_any(linenumber, 30, f = ceiling))
 56 | 
 57 | # aggregate by "linenumber" groups
 58 | all_books_byline <- all_books_long_10 %>% 
 59 |   group_by(book, linenumber, emotion) %>% 
 60 |   summarise(value = sum(value)) %>% 
 61 |   filter(value > 0) %>% 
 62 |   group_by(book, linenumber) %>% 
 63 |   mutate(test = value / sum(value) * n())
 64 | 
 65 | # add ids for each book
 66 | all_books_byline <- all_books_byline %>% group_by(book) %>% mutate(id = row_number())
 67 | 
 68 | # add unique identifier
 69 | all_books_byline_dt <- all_books_byline %>% mutate(book_line = paste0(book, linenumber))
 70 | all_books_byline_dt <- as.data.table(all_books_byline_dt)
 71 | 
 72 | # identify the 2 most common emotions in each linenumber group
 73 | all_books_byline_dt <- all_books_byline_dt[all_books_byline_dt[, .I[value >= max( value[value!=max(value)] )], by=book_line]$V1]
 74 | 
 75 | # apply an equal value to these emotions
 76 | all_books_byline_dt$value <- 1
 77 | 
 78 | # apply calculation to adjust appearance of the streamgraph and remove extreme spikes
 79 | all_books_byline_dt <- all_books_byline_dt %>% 
 80 |   group_by(book, emotion) %>% 
 81 |   mutate(test = n() / value)
 82 | 
 83 | # load Google font for visualisations
 84 | font_add_google("Noto Serif", "Noto Serif")
 85 | showtext_auto() 
 86 | 
 87 | # create custom colour palette
 88 | pal <- c("joy" = "#6b9a3e", 
 89 |          "trust" = "#25654d", 
 90 |          "surprise" = "#efd041", 
 91 |          "anticipation" = "#34478b", 
 92 |          "sadness" = "#4175a5", 
 93 |          "fear" = "#ed9953", 
 94 |          "anger" = "#9b262c", 
 95 |          "disgust" = "#807094")
 96 | 
 97 | # create plot for Sense & Sensibility
 98 | sense_p <- all_books_byline_dt %>% filter(book == "Sense & Sensibility") %>%
 99 |   ggplot(aes(x = id, y = test)) +
100 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
101 |   scale_fill_manual(values = pal) +
102 |   ggtitle("S E N S E   &   S E N S I B I L I T Y") +
103 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
104 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
105 |         panel.grid.major = element_blank(),
106 |         legend.background = element_blank(),
107 |         legend.key = element_blank(),
108 |         legend.text = element_text(size = 24),
109 |         legend.title = element_blank(),
110 |         legend.position = "none",
111 |         legend.box.margin = margin(r = 0, l = 0),
112 |         panel.grid.minor = element_blank(),
113 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
114 |         axis.text = element_blank(),
115 |         strip.background = element_blank(),
116 |         strip.text = element_text(color = "#69543f"),
117 |         axis.title = element_blank(),
118 |         axis.ticks = element_blank(),
119 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
120 | 
121 | # create plot for Emma
122 | emma_p <- all_books_byline_dt %>% filter(book == "Emma") %>%
123 |   ggplot(aes(x = id, y = test)) +
124 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
125 |   scale_fill_manual(values = pal) +
126 |   ggtitle("E M M A") +
127 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
128 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
129 |         panel.grid.major = element_blank(),
130 |         legend.background = element_blank(),
131 |         legend.text = element_text(size = 24),
132 |         legend.key = element_blank(),
133 |         legend.title = element_blank(),
134 |         legend.position = "none",
135 |         legend.box.margin = margin(r = 0, l = 0),
136 |         panel.grid.minor = element_blank(),
137 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
138 |         axis.text = element_blank(),
139 |         strip.background = element_blank(),
140 |         strip.text = element_text(color = "#69543f"),
141 |         axis.title = element_blank(),
142 |         axis.ticks = element_blank(),
143 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
144 | 
145 | # create plot for Northanger Abbey
146 | north_p <- all_books_byline_dt %>% filter(book == "Northanger Abbey") %>%
147 |   ggplot(aes(x = id, y = test)) +
148 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
149 |   scale_fill_manual(values = pal) +
150 |   ggtitle("N O R T H A N G E R   A B B E Y") +
151 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
152 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
153 |         panel.grid.major = element_blank(),
154 |         legend.background = element_blank(),
155 |         legend.text = element_text(size = 24),
156 |         legend.key = element_blank(),
157 |         legend.title = element_blank(),
158 |         legend.position = "none",
159 |         legend.box.margin = margin(r = 0, l = 0),
160 |         panel.grid.minor = element_blank(),
161 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
162 |         axis.text = element_blank(),
163 |         strip.background = element_blank(),
164 |         strip.text = element_text(color = "#69543f"),
165 |         axis.title = element_blank(),
166 |         axis.ticks = element_blank(),
167 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
168 | 
169 | # create plot for Mansfield Park
170 | mansfield_p <- all_books_byline_dt %>% filter(book == "Mansfield Park") %>%
171 |   ggplot(aes(x = id, y = test)) +
172 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
173 |   scale_fill_manual(values = pal) +
174 |   ggtitle("M A N S F I E L D   P A R K") +
175 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
176 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
177 |         panel.grid.major = element_blank(),
178 |         legend.background = element_blank(),
179 |         legend.key = element_blank(),
180 |         legend.text = element_text(size = 24),
181 |         legend.title = element_blank(),
182 |         legend.position = "none",
183 |         legend.box.margin = margin(r = 0, l = 0),
184 |         panel.grid.minor = element_blank(),
185 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
186 |         axis.text = element_blank(),
187 |         strip.background = element_blank(),
188 |         strip.text = element_text(color = "#69543f"),
189 |         axis.title = element_blank(),
190 |         axis.ticks = element_blank(),
191 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
192 | 
193 | # create plot for Pride & Prejudice
194 | pride_p <- all_books_byline_dt %>% filter(book == "Pride & Prejudice") %>%
195 |   ggplot(aes(x = id, y = test)) +
196 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
197 |   scale_fill_manual(values = pal) +
198 |   ggtitle("P R I D E   &   P R E J U D I C E") +
199 |   labs(caption = "Sentiment Analysis Data: NRC Word-Emotion Association Lexicon / saifmohammad.com | Visualisation: @filmicaesthetic") +
200 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
201 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
202 |         panel.grid.major = element_blank(),
203 |         legend.background = element_blank(),
204 |         legend.key = element_blank(),
205 |         legend.text = element_text(size = 18),
206 |         legend.title = element_blank(),
207 |         legend.position = "none",
208 |         legend.box.margin = margin(r = 0, l = 0),
209 |         panel.grid.minor = element_blank(),
210 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
211 |         axis.text = element_blank(),
212 |         strip.background = element_blank(),
213 |         strip.text = element_text(color = "#69543f"),
214 |         axis.title = element_blank(),
215 |         axis.ticks = element_blank(),
216 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
217 | 
218 | # build styled legend with a bar chart
219 | leg <- data.frame(EMOTION = as.factor(c("joy", "trust", "surprise", "anticipation", "sadness", "fear", "anger", "disgust")), value = c(1, 1, 1, 1, 1, 1, 1, 1))
220 | leg$EMOTION <- factor(leg$EMOTION, levels = leg$EMOTION)
221 | 
222 | # plot legend
223 | hist <- ggplot(leg, aes(x = EMOTION, y = value, fill = EMOTION)) + 
224 |   geom_bar(stat = "identity") +
225 |   geom_text(aes(y = 0.502, label = toupper(EMOTION)), angle = 90, family = "Noto Serif", color = c("#b07e13", "#f2ddb1", "#b07e13", "#f2ddb1", "#b07e13", "#b07e13", "#f2ddb1", "#b07e13"),  size = 8) +
226 |   geom_text(aes(y = 0.498, label = toupper(EMOTION)), angle = 90, family = "Noto Serif", color = c("#403020", "#b07e13", "#403020", "#b07e13", "#403020", "#403020", "#b07e13", "#403020"), size = 8) +
227 |   geom_text(aes(y = 0.95, label = "JANE AUSTEN"), family = "Noto Serif", color = "#f2ddb1", size = 1.5) +
228 |   geom_text(aes(y = 0.95001, label = "JANE AUSTEN"), family = "Noto Serif", color = "#291d12", size = 1.5) +
229 |   geom_text(aes(y = 0.05, label = "SIMON & SCHUSTER"), family = "Noto Serif", color = "#f2ddb1", size = 1.1) +
230 |   geom_text(aes(y = 0.05001, label = "SIMON & SCHUSTER"), family = "Noto Serif", color = "#291d12", size = 1.1) +
231 |   scale_fill_manual(values = pal) +
232 |   ggtitle(toupper("The Emotions of Jane Austen"), subtitle = "A visual interpretation of the emotion of language used in Jane Austen novels.") + 
233 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
234 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
235 |         panel.grid.major = element_blank(),
236 |         legend.background = element_blank(),
237 |         legend.key = element_blank(),
238 |         plot.margin = margin(r = 4, l = 4, unit = "cm"),
239 |         legend.position = "none",
240 |         legend.box.margin = margin(r = 0, l = 0),
241 |         panel.grid.minor = element_blank(),
242 |         plot.subtitle = element_text(hjust = 0.5, size= 18, margin = margin(t = 5, b = 5, unit = "pt")),
243 |         plot.title = element_text(angle = 0, family = "Noto Serif", color = "#69543f", size = 24, face="bold", hjust = 0.5, margin = margin(t = 12, unit = "pt")),
244 |         axis.text = element_blank(),
245 |         strip.background = element_blank(),
246 |         strip.text = element_text(color = "#80664d"),
247 |         axis.title = element_blank(),
248 |         axis.ticks = element_blank(),
249 |         text=element_text(family="Noto Serif", face = "bold", color = "#69543f"))
250 | 
251 | # combine all the plots into grid
252 | plot <- plot_grid(hist, sense_p, emma_p, north_p, mansfield_p, pride_p, nrow = 6, rel_heights = c(1.2, 1, 1, 1, 1, 1))
253 | 
254 | # save the plot
255 | ggsave("JaneAusten.png", plot = plot, dpi = 300, width = 13, height = 20, units = "cm")
256 | 


--------------------------------------------------------------------------------
/README.Rmd:
--------------------------------------------------------------------------------
  1 | ---
  2 | title: "The Emotions of Jane Austen"
  3 | output: 
  4 |   rmarkdown::github_document:
  5 |   theme: paper
  6 | ---
  7 | 
  8 | ```{r setup, include=FALSE}
  9 | knitr::opts_chunk$set(echo = TRUE)
 10 | 
 11 | # load packages
 12 | library(plyr)
 13 | library(stringr)
 14 | library(data.table)
 15 | library(ggplot2)
 16 | library(ggstream)
 17 | library(tidytext)
 18 | library(janeaustenr)
 19 | library(extrafont)
 20 | library(showtext)
 21 | library(dplyr)
 22 | library(gridExtra)
 23 | library(grid)
 24 | library(cowplot)
 25 | 
 26 | # load Google font for visualisations
 27 | font_add_google("Noto Serif", "Noto Serif")
 28 | showtext_auto() 
 29 | 
 30 | # format the data from {janeaustenr} for analysis
 31 | tidy_books <- austen_books() %>%
 32 |   group_by(book) %>%
 33 |   mutate(
 34 |     linenumber = row_number(),
 35 |     chapter = cumsum(str_detect(text, 
 36 |                                 regex("^chapter [\\divxlc]", 
 37 |                                       ignore_case = TRUE)))) %>%
 38 |   ungroup() %>%
 39 |   unnest_tokens(word, text)
 40 | 
 41 | # import NRC Emotion Lexicon
 42 | nrc_all <- get_sentiments("nrc")
 43 | 
 44 | # list of unique emotions
 45 | sentiments <- unique(nrc_all$sentiment)
 46 | 
 47 | # identify words that appear in the emotion lexicon
 48 | all_books <- tidy_books %>% 
 49 |   filter() %>%
 50 |   mutate(joy = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "joy"]),
 51 |          trust = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "trust"]),
 52 |          surprise = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "surprise"]),
 53 |          anticipation = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "anticipation"]),
 54 |          sadness = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "sadness"]),
 55 |          fear = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "fear"]),
 56 |          anger = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "anger"]),
 57 |          disgust = as.numeric(word %in% nrc_all$word[nrc_all$sentiment == "disgust"]))
 58 | 
 59 | # reshape the data frame
 60 | all_books_long <- melt(setDT(all_books), id.vars = c("book","linenumber","chapter","word"), variable.name = "emotion")
 61 | # round lines to nearest 30 (approximately 1 page) to help remove noise
 62 | all_books_long_10 <- all_books_long %>% mutate(linenumber = round_any(linenumber, 30, f = ceiling))
 63 | 
 64 | # aggregate by "linenumber" groups
 65 | all_books_byline <- all_books_long_10 %>% 
 66 |   group_by(book, linenumber, emotion) %>% 
 67 |   summarise(value = sum(value)) %>% 
 68 |   filter(value > 0) %>% 
 69 |   group_by(book, linenumber) %>% 
 70 |   mutate(test = value / sum(value) * n())
 71 | 
 72 | # add ids for each book
 73 | all_books_byline <- all_books_byline %>% group_by(book) %>% mutate(id = row_number())
 74 | 
 75 | # add unique identifier
 76 | all_books_byline_dt <- all_books_byline %>% mutate(book_line = paste0(book, linenumber))
 77 | all_books_byline_dt <- as.data.table(all_books_byline_dt)
 78 | 
 79 | # identify the 2 most common emotions in each linenumber group
 80 | all_books_byline_dt <- all_books_byline_dt[all_books_byline_dt[, .I[value >= max( value[value!=max(value)] )], by=book_line]$V1]
 81 | 
 82 | # apply an equal value to these emotions
 83 | all_books_byline_dt$value <- 1
 84 | 
 85 | # apply calculation to adjust appearance of the streamgraph and remove extreme spikes
 86 | all_books_byline_dt <- all_books_byline_dt %>% 
 87 |   group_by(book, emotion) %>% 
 88 |   mutate(test = n() / value)
 89 | 
 90 | # create custom colour palette
 91 | pal <- c("joy" = "#6b9a3e", 
 92 |          "trust" = "#25654d", 
 93 |          "surprise" = "#efd041", 
 94 |          "anticipation" = "#34478b", 
 95 |          "sadness" = "#4175a5", 
 96 |          "fear" = "#ed9953", 
 97 |          "anger" = "#9b262c", 
 98 |          "disgust" = "#807094")
 99 | 
100 | # create plot for Sense & Sensibility
101 | sense_p <- all_books_byline_dt %>% filter(book == "Sense & Sensibility") %>%
102 |   ggplot(aes(x = id, y = test)) +
103 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
104 |   scale_fill_manual(values = pal) +
105 |   ggtitle("S E N S E   &   S E N S I B I L I T Y") +
106 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
107 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
108 |         panel.grid.major = element_blank(),
109 |         legend.background = element_blank(),
110 |         legend.key = element_blank(),
111 |         legend.text = element_text(size = 24),
112 |         legend.title = element_blank(),
113 |         legend.position = "none",
114 |         legend.box.margin = margin(r = 0, l = 0),
115 |         panel.grid.minor = element_blank(),
116 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
117 |         axis.text = element_blank(),
118 |         strip.background = element_blank(),
119 |         strip.text = element_text(color = "#69543f"),
120 |         axis.title = element_blank(),
121 |         axis.ticks = element_blank(),
122 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
123 | 
124 | # create plot for Emma
125 | emma_p <- all_books_byline_dt %>% filter(book == "Emma") %>%
126 |   ggplot(aes(x = id, y = test)) +
127 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
128 |   scale_fill_manual(values = pal) +
129 |   ggtitle("E M M A") +
130 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
131 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
132 |         panel.grid.major = element_blank(),
133 |         legend.background = element_blank(),
134 |         legend.text = element_text(size = 24),
135 |         legend.key = element_blank(),
136 |         legend.title = element_blank(),
137 |         legend.position = "none",
138 |         legend.box.margin = margin(r = 0, l = 0),
139 |         panel.grid.minor = element_blank(),
140 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
141 |         axis.text = element_blank(),
142 |         strip.background = element_blank(),
143 |         strip.text = element_text(color = "#69543f"),
144 |         axis.title = element_blank(),
145 |         axis.ticks = element_blank(),
146 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
147 | 
148 | # create plot for Northanger Abbey
149 | north_p <- all_books_byline_dt %>% filter(book == "Northanger Abbey") %>%
150 |   ggplot(aes(x = id, y = test)) +
151 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
152 |   scale_fill_manual(values = pal) +
153 |   ggtitle("N O R T H A N G E R   A B B E Y") +
154 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
155 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
156 |         panel.grid.major = element_blank(),
157 |         legend.background = element_blank(),
158 |         legend.text = element_text(size = 24),
159 |         legend.key = element_blank(),
160 |         legend.title = element_blank(),
161 |         legend.position = "none",
162 |         legend.box.margin = margin(r = 0, l = 0),
163 |         panel.grid.minor = element_blank(),
164 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
165 |         axis.text = element_blank(),
166 |         strip.background = element_blank(),
167 |         strip.text = element_text(color = "#69543f"),
168 |         axis.title = element_blank(),
169 |         axis.ticks = element_blank(),
170 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
171 | 
172 | # create plot for Mansfield Park
173 | mansfield_p <- all_books_byline_dt %>% filter(book == "Mansfield Park") %>%
174 |   ggplot(aes(x = id, y = test)) +
175 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
176 |   scale_fill_manual(values = pal) +
177 |   ggtitle("M A N S F I E L D   P A R K") +
178 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
179 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
180 |         panel.grid.major = element_blank(),
181 |         legend.background = element_blank(),
182 |         legend.key = element_blank(),
183 |         legend.text = element_text(size = 24),
184 |         legend.title = element_blank(),
185 |         legend.position = "none",
186 |         legend.box.margin = margin(r = 0, l = 0),
187 |         panel.grid.minor = element_blank(),
188 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
189 |         axis.text = element_blank(),
190 |         strip.background = element_blank(),
191 |         strip.text = element_text(color = "#69543f"),
192 |         axis.title = element_blank(),
193 |         axis.ticks = element_blank(),
194 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
195 | 
196 | # create plot for Pride & Prejudice
197 | pride_p <- all_books_byline_dt %>% filter(book == "Pride & Prejudice") %>%
198 |   ggplot(aes(x = id, y = test)) +
199 |   geom_stream(bw = 0.6, n_grid = 16561, aes(fill = emotion), alpha = 0.96) +
200 |   scale_fill_manual(values = pal) +
201 |   ggtitle("P R I D E   &   P R E J U D I C E") +
202 |   labs(caption = "Sentiment Analysis Data: NRC Word-Emotion Association Lexicon / saifmohammad.com | Visualisation: @filmicaesthetic") +
203 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
204 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
205 |         panel.grid.major = element_blank(),
206 |         legend.background = element_blank(),
207 |         legend.key = element_blank(),
208 |         legend.text = element_text(size = 18),
209 |         legend.title = element_blank(),
210 |         legend.position = "none",
211 |         legend.box.margin = margin(r = 0, l = 0),
212 |         panel.grid.minor = element_blank(),
213 |         plot.title = element_text(color = "#69543f", size = 18, hjust = 0.5),
214 |         axis.text = element_blank(),
215 |         strip.background = element_blank(),
216 |         strip.text = element_text(color = "#69543f"),
217 |         axis.title = element_blank(),
218 |         axis.ticks = element_blank(),
219 |         text=element_text(family = "Noto Serif", face = "bold", color = "#69543f"))
220 | 
221 | # build styled legend with a bar chart
222 | leg <- data.frame(EMOTION = as.factor(c("joy", "trust", "surprise", "anticipation", "sadness", "fear", "anger", "disgust")), value = c(1, 1, 1, 1, 1, 1, 1, 1))
223 | leg$EMOTION <- factor(leg$EMOTION, levels = leg$EMOTION)
224 | 
225 | # plot legend
226 | hist <- ggplot(leg, aes(x = EMOTION, y = value, fill = EMOTION)) + 
227 |   geom_bar(stat = "identity") +
228 |   geom_text(aes(y = 0.502, label = toupper(EMOTION)), angle = 90, family = "Noto Serif", color = c("#b07e13", "#f2ddb1", "#b07e13", "#f2ddb1", "#b07e13", "#b07e13", "#f2ddb1", "#b07e13"),  size = 8) +
229 |   geom_text(aes(y = 0.498, label = toupper(EMOTION)), angle = 90, family = "Noto Serif", color = c("#403020", "#b07e13", "#403020", "#b07e13", "#403020", "#403020", "#b07e13", "#403020"), size = 8) +
230 |   geom_text(aes(y = 0.95, label = "JANE AUSTEN"), family = "Noto Serif", color = "#f2ddb1", size = 1.5) +
231 |   geom_text(aes(y = 0.95001, label = "JANE AUSTEN"), family = "Noto Serif", color = "#291d12", size = 1.5) +
232 |   geom_text(aes(y = 0.05, label = "SIMON & SCHUSTER"), family = "Noto Serif", color = "#f2ddb1", size = 1.1) +
233 |   geom_text(aes(y = 0.05001, label = "SIMON & SCHUSTER"), family = "Noto Serif", color = "#291d12", size = 1.1) +
234 |   scale_fill_manual(values = pal) +
235 |   ggtitle(toupper("The Emotions of Jane Austen"), subtitle = "A visual interpretation of the emotion of language used in Jane Austen novels.") + 
236 |   theme(panel.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
237 |         plot.background = element_rect(fill = "#f2f0ed", color = "#f2f0ed"),
238 |         panel.grid.major = element_blank(),
239 |         legend.background = element_blank(),
240 |         legend.key = element_blank(),
241 |         plot.margin = margin(r = 4, l = 4, unit = "cm"),
242 |         legend.position = "none",
243 |         legend.box.margin = margin(r = 0, l = 0),
244 |         panel.grid.minor = element_blank(),
245 |         plot.subtitle = element_text(hjust = 0.5, size= 18, margin = margin(t = 5, b = 5, unit = "pt")),
246 |         plot.title = element_text(angle = 0, family = "Noto Serif", color = "#69543f", size = 24, face="bold", hjust = 0.5, margin = margin(t = 12, unit = "pt")),
247 |         axis.text = element_blank(),
248 |         strip.background = element_blank(),
249 |         strip.text = element_text(color = "#80664d"),
250 |         axis.title = element_blank(),
251 |         axis.ticks = element_blank(),
252 |         text=element_text(family="Noto Serif", face = "bold", color = "#69543f"))
253 | 
254 | # combine all the plots into one
255 | plot <- plot_grid(hist, sense_p, emma_p, north_p, mansfield_p, pride_p, nrow = 6, rel_heights = c(1.2, 1, 1, 1, 1, 1))
256 | 
257 | ```
258 | 
259 | This personal project explores the literature of Jane Austen novels, provided by Julia Silge's <a href="https://github.com/juliasilge/janeaustenr">janeaustenr</a> package in R, using the <a href="https://saifmohammad.com/WebPages/NRC-Emotion-Lexicon.htm">NRC Word-Emotion Association Lexicon</a> by Saif Mohammad. The lexicon is used to identify words associated with a range of 8 emotions: joy, trust, surprise, anticipation, sadness, fear, anger, and disgust.
260 | 
261 | The project aims to capture a visual interpretation of the emotional journey in each novel with a single visualisation.
262 | 
263 | ``` {r main_plot, echo = FALSE, warning = FALSE, message = FALSE, fig.width = 13, fig.height = 20}
264 | plot
265 | ```
266 | 
267 | ## Files
268 | 
269 | 'the_emotions_of_jane_austen.R' : <a href="https://github.com/filmicaesthetic/JaneAustenStreamgraphs/blob/main/scripts/the_emotions_of_jane_austen.R">view code</a>
270 | 
271 | ## Process
272 | 
273 | * Load the NRC Word-Emotion Association Lexicon and Jane Austen novels datasets and prepare for further analysis.
274 | * Count occurrences of words with an emotional association for each line of each book.
275 | * Group lines approximately into pages (30 lines) to help reduce noise in the final visualisation.
276 | * Identify the two emotions with strongest association for each group.
277 | * Plot the results for each novel in a streamgraph.
278 | * Create grid of all plots and bookshelf-style legend.


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