├── README.md
└── PythonEDA.py


/README.md:
--------------------------------------------------------------------------------
 1 | # Python EDA on GDP per 1000 Hours Worked
 2 | 
 3 | This project explores global economic productivity through GDP per 1000 hours worked, with a focus on India and world comparisons.
 4 | 
 5 | ## Files Included
 6 | - `PythonEDA.py`: Python script for data analysis and visualization.
 7 | - `gdp_over_hours_worked.csv`: Dataset containing GDP and working hours data.
 8 | 
 9 | ## Features
10 | - Data cleaning and transformation
11 | - Visualizations (bar plots, scatter plots)
12 | - Regression modeling
13 | - Country-level comparisons (with emphasis on India)
14 | 
15 | ## How to Run
16 | Run the script using Python:
17 | ```bash
18 | python PythonEDA.py


--------------------------------------------------------------------------------
/PythonEDA.py:
--------------------------------------------------------------------------------
  1 | import pandas as pd
  2 | import numpy as np
  3 | import seaborn as sb
  4 | import matplotlib.pyplot as plt
  5 | from sklearn.linear_model import LinearRegression
  6 | from sklearn.model_selection import train_test_split
  7 | from sklearn.metrics import mean_squared_error, r2_score
  8 | from sklearn.preprocessing import StandardScaler
  9 | 
 10 | df = pd.read_csv("C:/Users/DELL/Desktop/4th Semester/Python Project/gdp_over_hours_worked.csv")
 11 | df['year'] = pd.to_datetime(df['year'], format='%Y')
 12 | 
 13 | # DATA CLEANING
 14 | print("--- Missing Values Before Cleaning ---")
 15 | print(df.isnull().sum())
 16 | 
 17 | num_cols = [
 18 |     'gdp', 'hours_worked', 'pop', 'gdp_over_k_hours_worked',
 19 |     'gdp_ppp_c', 'gdp_ppp', 'gdp_c', 'unemployment_r',
 20 |     'pop_over_65', 'working_age_pop_pct', 'employment_rate',
 21 |     'hours_per_employed', 'employed', 'total_hours_alternative',
 22 |     'gdp_ppp_over_k_hours_worked', 'gdp_over_pop',
 23 |     'gdp_ppp_over_pop', 'gdp_ppp_over_labor_force',
 24 |     'gdp_ppp_over_pop_c', 'gdp_over_pop_c', 'gdp_ppp_over_k_hours_worked_c'
 25 | ]
 26 | 
 27 | for col in num_cols:
 28 |     if df[col].dtype in ['float64', 'int64']:
 29 |         df[col].fillna(df[col].median(), inplace=True)
 30 | 
 31 | df['labor_force'].fillna(df['labor_force'].median(), inplace=True)
 32 | 
 33 | print("--- Missing Values After Cleaning ---")
 34 | print(df.isnull().sum())
 35 | 
 36 | #GLOBAL OBJECTIVE 1: World-Level Analysis
 37 | 
 38 | # Global Yearly Summary
 39 | global_yearly = df.groupby('year').agg({
 40 |     'gdp': 'sum',
 41 |     'pop': 'sum',
 42 |     'hours_worked': 'mean',
 43 |     'labor_force': 'sum'
 44 | }).reset_index()
 45 | global_yearly['gdp_per_capita'] = global_yearly['gdp'] / global_yearly['pop']
 46 | 
 47 | latest_year = df['year'].dt.year.max()
 48 | latest_data = df[df['year'].dt.year == latest_year]
 49 | top_gdp_per_hour = latest_data.sort_values(by='gdp_over_k_hours_worked', ascending=False).head(10)
 50 | bottom_gdp_per_hour = latest_data.sort_values(by='gdp_over_k_hours_worked').head(5)
 51 | 
 52 | top_gdp_rank = df.groupby(['year', 'country'])['gdp'].sum().reset_index()
 53 | 
 54 | if 'region' in df.columns:
 55 |     region_trend = df.groupby(['region', 'year'])['gdp_over_k_hours_worked'].mean().reset_index()
 56 | 
 57 | aging_corr = df[['pop_over_65', 'gdp', 'labor_force', 'hours_worked']].corr()
 58 | 
 59 | india = df[df['country'] == 'India']
 60 | 
 61 | #GLOBAL OBJECTIVE 2: Regression Model
 62 | reg_data = df[['gdp', 'hours_worked', 'pop', 'labor_force', 'gdp_over_k_hours_worked']]
 63 | X = reg_data[['gdp', 'hours_worked', 'pop', 'labor_force']]
 64 | y = reg_data['gdp_over_k_hours_worked']
 65 | 
 66 | scaler = StandardScaler()
 67 | X_scaled = scaler.fit_transform(X)
 68 | 
 69 | X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.2, random_state=42)
 70 | 
 71 | model = LinearRegression()
 72 | model.fit(X_train, y_train)
 73 | y_pred = model.predict(X_test)
 74 | 
 75 | print("\n--- Linear Regression Evaluation ---")
 76 | print("Mean Squared Error:", mean_squared_error(y_test, y_pred))
 77 | print("R-squared:", r2_score(y_test, y_pred))
 78 | 
 79 | plt.figure(figsize=(8, 6))
 80 | plt.scatter(y_test, y_pred, alpha=0.7, color='teal', marker = 'x')
 81 | plt.xlabel("Actual GDP per 1000 Hours")
 82 | plt.ylabel("Predicted GDP per 1000 Hours")
 83 | plt.title("Actual vs Predicted GDP/1000 Hours (Regression)")
 84 | plt.grid(True)
 85 | plt.tight_layout()
 86 | plt.show()
 87 | 
 88 | # GLOBAL OBJECTIVE 3: Efficiency Ranking
 89 | efficiency_df = df[df['year'].dt.year == 2021][['country', 'gdp_over_k_hours_worked']].dropna()
 90 | top_10 = efficiency_df.sort_values(by='gdp_over_k_hours_worked', ascending=False).head(10)
 91 | bottom_10 = efficiency_df.sort_values(by='gdp_over_k_hours_worked').head(10)
 92 | latest = df[df['year'].dt.year == 2021]
 93 | 
 94 | #Top-10 countries
 95 | plt.figure(figsize=(10, 5))
 96 | sb.barplot(data=top_10, y='country', x='gdp_over_k_hours_worked', palette='Greens')
 97 | plt.title('Top 10 Countries by GDP per 1000 Hours (2021)')
 98 | plt.xlabel('GDP per 1000 Hours')
 99 | plt.tight_layout()
100 | plt.show()
101 | 
102 | #Bottom-10 countries
103 | bottom_10 = (
104 |     latest[['country', 'gdp_over_k_hours_worked']]
105 |     .dropna()
106 |     .drop_duplicates(subset='country')
107 |     .sort_values(by='gdp_over_k_hours_worked')
108 |     .head(10)
109 | )
110 | 
111 | plt.figure(figsize=(10, 5))
112 | sb.barplot(data=bottom_10, x='country', y='gdp_over_k_hours_worked', palette='Reds')
113 | plt.title('Bottom 10 Countries by GDP per 1000 Hours (2021)')
114 | plt.xlabel('Country')
115 | plt.ylabel('GDP per 1000 Hours Worked')
116 | plt.xticks(rotation=45)
117 | plt.tight_layout()
118 | plt.show()
119 | 
120 | # GLOBAL OBJECTIVE 4: Developed vs Developing GDP per Capita
121 | dev = ['United States', 'Germany', 'UK', 'France', 'Japan', 'Canada']
122 | devg = ['India', 'Nigeria', 'Bangladesh', 'Pakistan', 'Kenya']
123 | 
124 | df['gdp_per_capita'] = df['gdp'] / df['pop']
125 | compare = df[df['country'].isin(dev + devg)].copy()
126 | compare['group'] = compare['country'].apply(lambda x: 'Developed' if x in dev else 'Developing')
127 | 
128 | plt.figure(figsize=(10, 5))
129 | sb.lineplot(data=compare, x='year', y='gdp_per_capita', hue='group', palette='Set2')
130 | plt.title('GDP per Capita: Developed vs Developing')
131 | plt.tight_layout()
132 | plt.show()
133 | 
134 | # OBJECTIVE 1: GDP Comparison (2021)
135 | selected = latest[latest['country'].isin(['India', 'China', 'United States', 'Germany'])]
136 | 
137 | plt.figure(figsize=(8, 6))
138 | sb.barplot(data=selected, x='country', y='gdp', palette='viridis')
139 | plt.title('GDP Comparison (2021)')
140 | plt.ylabel('GDP (USD)')
141 | plt.xlabel('Country')
142 | plt.tight_layout()
143 | plt.show()
144 | 
145 | # OBJECTIVE 2: India GDP Trend Over Time
146 | plt.figure(figsize=(10, 5))
147 | plt.plot(india['year'], india['gdp'], label='GDP', color='skyblue', linewidth=2)
148 | plt.scatter(india['year'], india['gdp'], color='navy', s=50)
149 | plt.title('India GDP Over the Years')
150 | plt.xlabel('Year')
151 | plt.ylabel('GDP')
152 | plt.grid(True)
153 | plt.legend()
154 | plt.tight_layout()
155 | plt.show()
156 | 
157 | # OBJECTIVE 3: Correlation Heatmaps (World & India)
158 | # Global Correlation Heatmap
159 | corr = df[['gdp', 'hours_worked', 'gdp_over_k_hours_worked', 'pop']].corr()
160 | plt.figure(figsize=(8, 6))
161 | sb.heatmap(corr, annot=True, cmap='coolwarm', linewidths=0.5)
162 | plt.title('Correlation Matrix (World)')
163 | plt.tight_layout()
164 | plt.show()
165 | 
166 | # 🇮🇳 India Correlation Heatmap
167 | india_numeric = india.select_dtypes(include=['float64', 'int64'])
168 | valid_cols = india_numeric.loc[:, india_numeric.nunique() > 1].dropna(axis=1, how='any')
169 | corr_india = valid_cols.corr()
170 | 
171 | plt.figure(figsize=(10, 6))
172 | sb.heatmap(corr_india, annot=True, cmap='coolwarm', linewidths=0.5, vmin=-1, vmax=1)
173 | plt.title('Correlation Matrix (India)')
174 | plt.tight_layout()
175 | plt.show()
176 | 
177 | # OBJECTIVE 5: India's Share of Global GDP Over Time
178 | total_gdp_year = df.groupby('year')['gdp'].sum().reset_index(name='global_gdp')
179 | india_gdp = india[['year', 'gdp']].merge(total_gdp_year, on='year')
180 | india_gdp['share'] = india_gdp['gdp'] / india_gdp['global_gdp'] * 100
181 | 
182 | plt.figure(figsize=(10, 5))
183 | plt.plot(india_gdp['year'], india_gdp['share'], color='purple')
184 | plt.title("India's Share of Global GDP Over Time (%)")
185 | plt.xlabel('Year')
186 | plt.ylabel('Share (%)')
187 | plt.grid(True)
188 | plt.tight_layout()
189 | plt.show()
190 | 
191 | # OBJECTIVE 6: India - unemployment rate (since employment rate is flat)
192 | plt.figure(figsize=(10, 5))
193 | plt.plot(india['year'], india['unemployment_r'] * 100, label='Unemployment Rate (%)', color='red', linewidth=2)
194 | plt.title('India: Unemployment Rate Over Time')
195 | plt.xlabel('Year')
196 | plt.ylabel('Percentage')
197 | plt.grid(True)
198 | plt.legend()
199 | plt.tight_layout()
200 | plt.show()
201 | 


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