AQI Prediction System using XGBoost

An AI-powered Air Quality Index (AQI) prediction and monitoring system developed as an MCA final year project.
This application predicts AQI levels using Machine Learning models and provides real-time air quality insights through an interactive web dashboard.

---

Project Overview

The AQI Prediction System is designed to analyze pollutant data and predict air quality categories and AQI values using XGBoost Machine Learning models.
The system also integrates live pollution data APIs and visual dashboards for better environmental awareness and decision-making.

---

Key Features

• Real-time AQI monitoring
• AQI value prediction using XGBoost Regression
• AQI category classification
• SHAP explainability integration
• Live pollution data integration using OpenWeather API
• Interactive and responsive dashboard
• AQI trend visualization
• Health advisory based on AQI levels
• Comparison between predicted AQI and live AQI
• Clean and modern user interface

---

Machine Learning Features

• Data preprocessing and cleaning
• SMOTEENN for handling class imbalance
• XGBoost Regression Model
• XGBoost Classification Model
• Model evaluation and prediction analysis
• SHAP Explainable AI visualization

---

Technology Stack

Frontend

• HTML5
• CSS3
• JavaScript
• Bootstrap

Backend

• Python
• Flask

Machine Learning

• XGBoost
• Scikit-learn
• Pandas
• NumPy
• SHAP

APIs

• OpenWeather Air Pollution API

---

Project Structure

AQI-Prediction-System/
│
├── data/                  # Dataset files
├── logs/                  # Log files
├── models/                # Trained ML models
├── screenshots/           # Project screenshots
├── static/                # CSS, JS, Images
├── templates/             # HTML templates
├── app.py                 # Main Flask application
├── architecture-diagram.png
├── README.md
├── requirements.txt
└── .gitignore

---

Installation

Clone the repository

git clone https://github.com/jeesonjustin/AQI-Prediction-System.git

Navigate to project folder

cd AQI-Prediction-System

Create virtual environment

python -m venv .venv

Activate virtual environment

.venv\Scripts\activate

Install dependencies

pip install -r requirements.txt

Run the Project

python app.py

Open browser and visit:

http://127.0.0.1:5000

---

AQI Categories

AQI Range	Category
0 – 50	Good
51 – 100	Satisfactory
101 – 200	Moderate
201 – 300	Poor
301 – 400	Very Poor
401 – 500	Severe

---

UserInterface

---

Future Enhancements

• Mobile application integration
• Advanced forecasting models
• User authentication system
• Historical AQI analytics
• Multi-city comparison dashboard

---

Academic Information

Project Title: AQI-Prediction-System

Course: Master of Computer Applications (MCA)

Institution: SCMS School of Engineering and Technology

University: APJ Abdul Kalam Technological University

---

Developed By

Jeeson Justin MCA Student | UI/UX Designer | Full Stack Developer

• GitHub: https://github.com/jeesonjustin
• LinkedIn: https://linkedin.com/in/jeesonjustin

---