🌍 Air Quality Intelligence Dashboard

A machine learning-powered dashboard for analyzing, predicting, and exploring Air Quality Index (AQI) trends across major Indian cities.

Built using Python, XGBoost, Streamlit, Plotly, and Scikit-Learn, the project helps users understand air pollution patterns, identify key AQI drivers, and simulate environmental interventions through interactive visualizations.

---

📌 Project Overview

Air pollution is one of the most critical environmental challenges affecting public health in urban areas.

This project aims to:

• Analyze historical AQI and pollutant data from Indian cities.
• Build a machine learning model to predict AQI.
• Identify the most influential pollution factors.
• Provide interactive visualizations and insights.
• Simulate "what-if" pollution scenarios to estimate AQI changes.

---

🚀 Features

📈 City Insights

• Select any available city.

• View predicted AQI using the trained ML model.

• AQI gauge visualization.

• AQI category classification:

  • Good
  • Satisfactory
  • Moderate
  • Poor
  • Very Poor
  • Severe

• 90-day AQI trend analysis.

🔍 AQI Drivers Analysis

• Feature importance visualization.
• Identification of major pollutants affecting AQI.
• PM2.5 sensitivity analysis.
• Interactive plots for model interpretation.

🧪 Scenario Simulator

Users can simulate pollution control measures by adjusting:

• PM2.5 concentration
• PM10 concentration
• NO₂ concentration

The dashboard then predicts:

• Updated AQI
• AQI category
• Health advisory

---

🧠 Machine Learning Pipeline

Data Preprocessing

• Removed records with missing AQI values.
• Filled missing pollutant values using city-wise median imputation.
• Encoded city names using Label Encoding.
• Created cleaned dataset for modeling.

Models Evaluated

• Linear Regression
• Random Forest Regressor
• XGBoost Regressor

Final Model

✅ XGBoost Regressor

Selected based on superior predictive performance.

---

📊 Exploratory Data Analysis (EDA)

The project includes:

EDA Phase 1

• Data cleaning
• Missing value handling
• Pollutant distribution analysis
• City-wise AQI exploration

EDA Phase 2

• Model comparison
• Feature importance analysis
• Partial dependence style AQI sensitivity studies

---

🛠️ Tech Stack

Programming Language

• Python

Data Analysis

• Pandas
• NumPy

Visualization

• Matplotlib
• Seaborn
• Plotly

Machine Learning

• Scikit-Learn
• XGBoost

Web Application

• Streamlit

---

📂 Project Structure

air-quality-dashboard/
│
├── data/
│   ├── city_day.csv
│   └── city_day_clean.csv
│
├── model/
│   └── xgb_model.pkl
│
├── plots/
│
├── notebooks/
│   ├── 01_eda.ipynb
│   └── 02_eda.ipynb
│
├── app.py
├── utils.py
├── requirements.txt
└── README.md

---

⚙️ Installation

Clone the repository:

git clone https://github.com/yourusername/air-quality-dashboard.git
cd air-quality-dashboard

Install dependencies:

pip install -r requirements.txt

Run the Streamlit application:

streamlit run app.py

---

📈 Dashboard Preview

City Insights

• AQI Prediction
• AQI Gauge
• Historical AQI Trends

AQI Drivers

• Feature Importance Ranking
• PM2.5 Impact Analysis

Scenario Analysis

• Pollution Reduction Simulation
• AQI Forecasting

---

📚 Dataset

Indian city air quality dataset containing:

• AQI
• PM2.5
• PM10
• NO₂
• SO₂
• CO
• O₃
• City information
• Date information

---

🌐 Live Demo

Try the deployed application here:

👉 https://airaqi.streamlit.app/

Explore:

• AQI Prediction Dashboard
• City-wise AQI Insights
• AQI Driver Analysis
• Pollution Scenario Simulator