Aadhaar Data Analysis

Analysis of Aadhaar enrolment and update patterns across India to identify meaningful trends and insights.

Problem Statement

Unlocking Societal Trends in Aadhaar Enrolment and Updates - Identify meaningful patterns, trends, anomalies, or predictive indicators in Aadhaar data and translate them into clear insights or solution frameworks that can support informed decision-making and system improvements.

Datasets

This analysis uses three types of Aadhaar data:

1. Demographic Updates (demo_combined.csv)

Count of demographic information updates by age group.

2. Biometric Updates (bio_combined.csv)

Count of biometric (fingerprint/iris) updates by age group.

3. Enrolments (enroll_combined.csv)

Count of new Aadhaar enrolments by age group.

4. Geographic Boundaries (INDIA_STATES.geojson)

GeoJSON file containing state boundary polygons for map visualizations.

Data Dictionary

Column	Type	Description
date	datetime	Date of activity (DD-MM-YYYY format)
state	string	State name (lowercase)
district	string	District name
pincode	integer	6-digit postal code
demo_age_5_17	integer	Demographic updates for ages 5-17
demo_age_17_	integer	Demographic updates for ages 17+
bio_age_5_17	integer	Biometric updates for ages 5-17
bio_age_17_	integer	Biometric updates for ages 17+
age_0_5	integer	New enrolments for ages 0-5
age_5_17	integer	New enrolments for ages 5-17
age_18_greater	integer	New enrolments for ages 18+

Setup Instructions

Prerequisites

• Python 3.8 or higher
• pip package manager

Installation

1. Clone or download this repository

2. Install dependencies:

   pip install -r requirements.txt

3. Set up data path (optional - defaults to /content/drive/MyDrive/uidai for Google Colab):

   # Linux/Mac
   export AADHAAR_DATA_PATH="/path/to/your/data"
   
   # Windows (PowerShell)
   $env:AADHAAR_DATA_PATH="C:\path\to\your\data"
   
   # Windows (CMD)
   set AADHAAR_DATA_PATH=C:\path\to\your\data

4. Organize your data:

   your_data_path/
   ├── combined_dataset2/
   │   ├── demo_combined.csv
   │   ├── bio_combined.csv
   │   └── enroll_combined.csv
   └── INDIA_STATES.geojson
   

5. Run the analysis:

   jupyter notebook

Project Structure

aadhaar-analysis/
├── config.py                          # Configuration and paths
├── constants.py                       # Constants and definitions
├── requirements.txt                   # Python dependencies
├── README.md                          # This file
├── utils/                            # Utility modules
│   ├── __init__.py
│   ├── data_loader.py                # Data loading functions
│   ├── preprocessing.py              # Data cleaning & merging
│   ├── features.py                   # Feature engineering
│   ├── visualizations.py             # Plotting functions
│   └── ml_models.py                  # Machine learning models
└── analysis.ipynb                # Analysis notebook

Usage

Using Utility Functions

# Import utilities
from utils import (
    load_aadhaar_data,
    format_dates,
    merge_datasets,
    clean_merged_data,
    add_all_features,
    plot_state_map,
    plot_bar
)

# Load and prepare data
demo, bio, enroll = load_aadhaar_data()
demo = format_dates(demo)
bio = format_dates(bio)
enroll = format_dates(enroll)

# Merge and clean
df = merge_datasets(demo, bio, enroll)
df = clean_merged_data(df)

# Add features
df = add_all_features(df)

# Visualize
state_totals = df.groupby('state')['total_activity'].sum().reset_index()
plot_state_map(state_totals, 'total_activity', 'Total Activity by State')

Machine Learning

from utils import train_all_models

# Train all ML models
results = train_all_models(df)

# Access results
print(f"Regression R²: {results['regression']['metrics']['r2_test']:.4f}")
print(f"Classification Accuracy: {results['classification']['metrics']['accuracy_test']:.4f}")
print(f"Clustering Silhouette: {results['clustering']['metrics']['silhouette']:.4f}")

Visualizations

The analysis includes:

1. Choropleth Maps - State-wise activity distribution
2. Bar Charts - Top states by activity type
3. Time Series - Daily activity trends
4. Heatmaps - Temporal patterns
5. Scatter Plots - District-level analysis

Machine Learning Models

The analysis includes three ML models:

1. Linear Regression

• Purpose: Predict total activity based on temporal features
• Features: day, month, is_weekend
• Target: total_activity

2. Logistic Regression

• Purpose: Classify dominant activity type
• Features: demo_ratio, bio_ratio, enrol_ratio, day_of_week, month
• Target: activity_type

3. K-Means Clustering

• Purpose: Group similar activity patterns
• Features: Standardized activity ratios and temporal features

Testing

To test the utility functions:

# Test imports
python -c "from utils import *; print('All imports successful')"

---

Last Updated: January 14, 2026