📊 Crime Risk Scoring System

A machine learning project to analyze district-wise crime data (2010–2020) and classify regions into risk levels 🚨

---

📌 About the Project

This project analyzes crime data across Indian districts and assigns a risk score based on different types of crimes.
The workflow includes:

• Collecting raw district-wise crime data (2010–2020) 🗂️
• Assigning weights to different crime categories (e.g., murder > theft) ⚖️
• Normalizing by population → per capita crime 📐
• Scaling by median values for comparability 📊
• Computing risk scores for each district
• Training a Logistic Regression Model to classify into risk levels (0 = Safe → 4 = Very High Risk)

---

🛠️ Tech Stack

• Python 🐍 – Data analysis & ML
• Pandas – Data processing
• NumPy – Numerical operations
• Matplotlib / Seaborn – Visualization
• Scikit-learn – ML model (Logistic Regression)

---

📂 Dataset

• Source: District-wise crimes committed (IPC), India (2010–2020)
• Data contains:
  • Yearly district-wise crime counts
  • Population data
  • Features created:
    • Per capita crime rates
    • Weighted scores based on severity
• Files are stored in: data/ 📁

---

🤖 Machine Learning Model

• Model used: Logistic Regression
• Why Logistic Regression?
  • ✅ Simple & interpretable
  • ✅ Handles categorical risk levels (0–4)
  • ✅ Easy to explain coefficients
• Why not SVM / Random Forest?
  • ❌ SVM → better for complex, non-linear data, harder to train, not multinomial by default
  • ❌ Random Forest → powerful but computationally heavy & less interpretable

---

⚡ How to Run

# Clone the repository
git clone https://github.com/your-username/crime-risk-scoring.git
cd crime-risk-scoring

# Create virtual environment (optional)
python -m venv venv
source venv/bin/activate   # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Run the notebook / script
jupyter notebook analysis.ipynb