Comprehensive Analysis and Forecasting of Air Quality in Kathmandu

An end-to-end data science project that analyzes historical air quality trends in Kathmandu, predicts AQI using machine learning, and provides real-time forecasting using Facebook Prophet — all wrapped in an interactive Streamlit dashboard.

🔗 Live App: kathmanduairqualityforecasting.streamlit.app

---

Table of Contents

• Table of Contents
• Project Overview
• Objectives
• Features
• Methodology
• Models Used
• Dashboard Preview
• Technologies Used
• Folder Structure
• Run Locally
• Results
• References

---

Project Overview

Air pollution in Kathmandu has become a growing public health concern. This project provides a comprehensive analysis and predictive framework for monitoring and forecasting the Air Quality Index (AQI) using real-world environmental and meteorological data.

Using statistical modeling and machine learning, we built:

• A predictive AQI model based on environmental inputs.
• A time-series forecasting model for future AQI estimation.
• An interactive Streamlit dashboard for visualization, prediction, and forecasting.

---

Objectives

• Analyze historical AQI and weather data of Kathmandu.
• Forecast short-term AQI trends (24, 48, 72 hours).
• Develop a Random Forest-based model for real-time AQI prediction.
• Deploy an interactive web dashboard for public and policymaker use.

---

Features

• Visualize: Interactive charts (line, bar, heatmap, pie) showing AQI trends, statistics, and category breakdowns.
• Predict: Input current environmental parameters to get predicted AQI and health advisories.
• Forecast: Generate 24/48/72-hour AQI forecasts using Facebook Prophet.
• Live Dashboard: Deployed and accessible online.

---

Methodology

1. Data Collection: Open-Meteo API + government air quality datasets.
2. Preprocessing: Handled missing values, normalized features, converted time series.
3. Exploratory Analysis: Visualized trends, correlations, and seasonal effects.
4. Modeling:
   • Random Forest Regressor for real-time AQI prediction.
   • Facebook Prophet for time-series forecasting.
5. Deployment: Streamlit Cloud

---

Models Used

Model	Purpose	Metric (R² Score)
Random Forest Regressor	Predict AQI from environment inputs	0.91
Facebook Prophet	Forecast AQI (24–72 hrs)	N/A (Time series)

---

Dashboard Preview

---

Technologies Used

• Python
• Streamlit (Dashboard)
• Facebook Prophet (Time-series forecasting)
• Scikit-learn (Machine Learning)
• Pandas, NumPy, Matplotlib, Seaborn, Plotly (EDA & Visualization)

---

Sample Folder Structure

├── Dashboard/
│   ├── app.py                   # Main Streamlit app
│   ├── models/
│   │   └── model.pkl            # Trained Random Forest model
│   ├── data/
│   │   └── air_quality_data.csv # Historical dataset
│   └── models/
│       └── prophet_model.py     # Forecasting functions

---

Run Locally

1. Clone the repo

git clone https://github.com/poudelsangam77/Comprehensive-Analysis-and-Forecasting-of-Air-Quality-in-Kathmandu.git
cd kathmandu-air-quality-forecasting/Dashboard

2. Create virtual environment & activate

python -m venv venv
source venv/bin/activate  # on Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Run the app

streamlit run app.py

---

Results

• AQI prediction achieved R² = 0.91 with Random Forest Regressor.
• Seasonal trends and daily pollution cycles clearly visualized.
• Streamlit dashboard enabled public-friendly access to air quality insights and forecasts.

---

References

• Facebook Prophet Docs
• Streamlit Docs
• Open-Meteo API
• OpenAQ API

---

Developed by
Sangam Paudel · Saroj Rawal · Subesh Yadav
Department of Electronics and Computer Engineering, Pulchowk Campus
Tribhuvan University, Nepal