🗳️ Election Party Predictor

A Machine Learning based Election Prediction System built using Python, Scikit-Learn, CatBoost, XGBoost, and Streamlit.

This project predicts the leading political party based on constituency and state-level election information using multiple machine learning classification algorithms.

---

🚀 Features

• 📊 Election Dataset Analysis
• 🤖 Multiple Machine Learning Models
• 📈 Interactive Visualizations
• 🧩 Confusion Matrix Visualization
• 🔍 Real-Time Election Party Prediction
• 🌐 Streamlit Interactive Dashboard
• ⚡ Model Accuracy Comparison
• 📄 Classification Report Display

---

🛠️ Technologies Used

Programming Language

• Python

Libraries & Frameworks

• Pandas
• NumPy
• Matplotlib
• Scikit-Learn
• CatBoost
• XGBoost
• Streamlit

---

📁 Dataset Information

The dataset contains:

Column	Description
State/UT	State or Union Territory
Constituency	Election Constituency
Const. No.	Constituency Number
Leading Party	Winning Political Party
Trailing Party	Runner-Up Political Party
Margin	Winning Margin

---

⚙️ Data Preprocessing

The project performs:

• Missing Value Handling
• Feature Selection
• Data Cleaning
• Categorical Encoding
• Numerical Standardization
• Pipeline-based Data Transformation

---

🤖 Machine Learning Models Used

Model	Purpose
RandomForestClassifier	Main Prediction Model
DecisionTreeClassifier	Baseline Model
GradientBoostingClassifier	Ensemble Boosting Model
CatBoostClassifier	Advanced Categorical ML Model

---

📊 Model Performance

Model	Accuracy
Random Forest	~58%
Gradient Boosting	~58%
Decision Tree	~56%
CatBoost	Comparable Performance

Accuracy may vary slightly depending on train-test split.

---

🖥️ Streamlit Dashboard Features

The application includes:

• Interactive Prediction UI
• Dynamic State Selection
• Constituency-Based Prediction
• Dataset Overview Dashboard
• Model Accuracy Graph
• Constituency Distribution Graph
• Confusion Matrix Visualization
• Classification Report Table

---

📸 Project Preview

Dashboard Includes

• Election Prediction Interface
• ML Model Comparison
• Dataset Insights
• Interactive Visualizations
• Prediction Results

---

📂 Project Structure

Election-Party-Predictor/
│
├── app.py
├── Model.py
├── Election_Dataset.csv
├── requirements.txt
├── README.md
└── predictions.csv

---

⚙️ Installation & Setup

1️⃣ Clone Repository

git clone https://github.com/your-username/Election-Party-Predictor.git

---

2️⃣ Move into Project Directory

cd Election-Party-Predictor

---

3️⃣ Install Dependencies

pip install -r requirements.txt

---

4️⃣ Run Streamlit Application

streamlit run app.py

---

📦 Requirements

streamlit
pandas
numpy
matplotlib
scikit-learn
catboost
xgboost

---

📈 Workflow

Dataset Loading
       ↓
Data Cleaning
       ↓
Feature Engineering
       ↓
Train-Test Split
       ↓
Preprocessing Pipeline
       ↓
Model Training
       ↓
Model Evaluation
       ↓
Prediction System
       ↓
Streamlit Deployment

---

🔍 Prediction System

The application predicts the most likely leading political party using:

• State/UT
• Constituency
• Constituency Number

as input features.

---

📊 Evaluation Metrics Used

• Accuracy Score
• Error Rate
• Confusion Matrix
• Classification Report

---

🌟 Key Learning Outcomes

This project demonstrates practical understanding of:

• Data Science Workflow
• Machine Learning Pipelines
• Feature Engineering
• Ensemble Learning
• Classification Algorithms
• Data Visualization
• Streamlit Deployment
• Model Evaluation

---

🚀 Future Improvements

• Add Historical Election Data
• Add Demographic Features
• Add Vote Share Analysis
• Improve Prediction Accuracy
• Add Deep Learning Models
• Deploy on Streamlit Cloud
• Add Real-Time Analytics

---

👨‍💻 Author

Arpit Shirbhate

---

⭐ Support

If you like this project, consider giving it a ⭐ on GitHub.

---

📜 License

This project is open-source and available under the MIT License.