Skip to content

Repository files navigation

🌍 N.O.V.A. — Geospatial Ozone Predictor

An advanced atmospheric intelligence platform using NASA data (1994–2021), Stochastic Machine Learning, and Real-Time Telemetry.

NOVA Dashboard


📌 Overview

N.O.V.A. (North American Ozone Visual Analytics) is a Mission Control–style atmospheric intelligence system designed to analyze, visualize, and predict stratospheric and tropospheric Ozone (O₃) concentrations across Western North America.

It transforms nearly 30 years of NASA atmospheric back-trajectory data into an interactive desktop dashboard combining:

  • 📊 Historical Geospatial Visualization
  • 🤖 Random Forest Machine Learning Prediction
  • 🛰️ Real-Time ISS Telemetry Tracking

✨ Core Features

Feature Description
🌌 3D Geospatial Visualization Interactive 3D atmospheric ozone distribution
📈 Trend Analysis Historical line graph and seasonal analysis
🤖 AI Prediction Engine 50-tree Random Forest Regressor
🛰️ Live ISS Tracker Real-time satellite telemetry via API
🎛️ Mission Control UI Dark cyber-corporate themed interface

🖼 UI Preview

1️⃣ Dashboard


2️⃣ 3D Visualization


3️⃣ Line Graph Analysis


4️⃣ AI Prediction Engine


5️⃣ Live Satellite Location (ISS Tracking)


🧠 Data Science & AI Logic

📥 Independent Variables (Features)

  • Pressure (hPa) — Represents altitude
  • Latitude
  • Longitude
  • Month
  • Year

📤 Dependent Variable (Target)

  • Ozone Concentration (ppbv)

🤖 Model Details

  • Algorithm: Random Forest Regressor
  • Trees: 50 (n_estimators=50)
  • Train/Test Split: 80/20
  • Typical R² Score: 0.85 – 0.94
  • Most Influential Feature: Pressure (Altitude)

Random Forest was selected because atmospheric ozone behavior is non-linear and seasonal, making linear regression insufficient.


🛰️ Real-Time Telemetry

The ISS tracking system:

  • Fetches live coordinates from a public API
  • Updates every 3 seconds
  • Runs in a background thread
  • Prevents UI freezing using multithreading

🏗 Architecture

N.O.V.A. follows the Model–View–Controller (MVC) pattern:

Layer Role
Model Data processing & AI engine
View Tkinter GUI interface
Controller User-triggered simulation & visualization logic

🛠 Installation & Setup

🔹 1. Clone Repository

git clone https://github.com/arshc0der/N.O.V.A-Geospatial-Ozone-Predictor.git
cd N.O.V.A-Geospatial-Ozone-Predictor

🔹 2. Create Virtual Environment (Recommended)

python -m venv venv

Windows

venv\Scripts\activate

macOS / Linux

source venv/bin/activate

🔹 3. Install Dependencies

pip install -r requirements.txt

🔹 4. Ensure Dataset File Exists

Place this file in the root directory:

Receptor_western_NAmerica_ozone_obs_1994_2021_from900to300.csv

🔹 5. Run Application

python app.py

📦 requirements.txt

pandas
numpy
scikit-learn
matplotlib
seaborn
requests
tkintermapview

🤝 Contributing

  1. Fork the repository
  2. Create a feature branch
  3. Commit changes
  4. Submit a Pull Request

📜 License

Distributed under the MIT License. © 2026 Arsh


🚀 Built using NASA atmospheric back-trajectory data.

About

An AI-powered geospatial intelligence dashboard for predicting atmospheric ozone levels using 27 years of NASA data. Features 3D climate mapping and live satellite tracking.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages