🌍 Air Pollution Data Analysis Project

📌 Project Overview

This project analyzes global air quality data using Python, Pandas, NumPy, Matplotlib, and Seaborn.
It explores pollution levels across cities, evaluates pollutant distributions, detects anomalies, and checks if temperature or season impacts pollution.
The project also estimates a simple Air Quality Index (AQI) for different cities.

---

⚙️ Technologies Used

• Python 🐍
• Pandas (Data Handling)
• NumPy (Numerical Operations)
• Matplotlib & Seaborn (Visualization)
• Plotly (Interactive Visuals)

---

📂 Dataset

• File: global_air_quality_data.csv
• Columns:
  • City, Country
  • Date
  • PM2.5, PM10, NO2, SO2, CO, O3
  • Temperature, Humidity, Wind Speed

---

🔑 Features Implemented

1. Data Cleaning & Preprocessing

   • Duplicate removal
   • Date conversion to datetime
   • Handling missing values

2. Feature Engineering

   • Extracted Year, Month, Season, WeekDay
   • Rolling average for PM2.5
   • Z-score based anomaly detection

3. Exploratory Data Analysis (EDA)

   • Average pollution level by city
   • Monthly & seasonal pollution trends
   • Correlation heatmap (pollutants + weather factors)
   • Temperature vs PM2.5 (scatter & regression)
   • Top 5 most polluted cities
   • Distribution of PM2.5
   • Worst pollution day per city

4. AQI Estimation

   • Simple AQI = mean of pollutants
   • Ranked cities by AQI
   • Identified best city (least polluted)

---

📊 Visualizations

1. Monthly PM2.5 Trend

2. Temperature vs PM2.5 Scatter

3. Correlation Heatmap

4. Top 5 Polluted Cities

5. PM2.5 Distribution

---

📌 Key Insights (from dataset)

1. Seasonal impact is negligible – Month-wise PM2.5 variation is very small (range ≈ 5.03 on mean 77.45; ~6.5%).
2. Temperature has no real effect – PM2.5 vs Temperature correlation is nearly zero (r ≈ -0.012, n=10,000).
3. Overall pollution levels are high – Average PM2.5 ≈ 77.45, which is 3.1× higher than the WHO safe limit (25).

---

🚀 How to Run

1. Clone this repo or copy the code into a Jupyter Notebook / Google Colab.
2. Place the dataset global_air_quality_data.csv in the same folder.
3. Run the notebook step by step.
4. Output:
   • Cleaned dataset
   • CSV reports (city averages, AQI, seasonal data, anomalies)
   • PNG charts (saved inside visual/ folder)

---

✅ Future Improvements

• Implement official AQI formula (EPA/WHO breakpoints).
• Build predictive models (forecast pollution using weather).
• Add interactive dashboards (Plotly/Streamlit).

---

👤 Author

• Madhav Ganorkar
• Tools: Python, Pandas, NumPy, Matplotlib, Seaborn
• Contact: LinkedIn