A World Away: Hunting for Exoplanets
NASA Space Apps Challenge 2025
🌌 Project Overview
Our project leverages the power of artificial intelligence and machine learning to detect and classify exoplanets using data from NASA's premier space missions. We've developed sophisticated models that analyze observational data to identify potential exoplanets and classify them with high accuracy.
By combining data from multiple NASA missions including Kepler, TESS, and K2, our system provides a comprehensive approach to exoplanet detection, making it easier for researchers and enthusiasts to explore worlds beyond our solar system.
🛰️ Mission Data Sources
Kepler Mission
NASA's Kepler mission revolutionized exoplanet discovery by monitoring over 150,000 stars simultaneously. Our model analyzes Kepler Objects of Interest (KOI) to identify potential exoplanet candidates based on transit photometry data.
TESS Mission
The Transiting Exoplanet Survey Satellite (TESS) continues the legacy of exoplanet discovery by surveying the entire sky. Our TESS model processes TESS Objects of Interest (TOI) to detect planets around nearby stars.
K2 Mission
K2, the extended mission of Kepler, explores diverse regions of the sky. Our integrated approach incorporates K2 data to provide comprehensive exoplanet detection capabilities.
✨ Key Features
🤖 AI-Powered Detection
Advanced Random Forest algorithms trained on real NASA mission data for accurate exoplanet classification.
📊 Batch Processing
Upload CSV or JSON files to process multiple observations simultaneously with instant results.
🎯 High Accuracy
Our models achieve exceptional accuracy in distinguishing between confirmed planets, candidates, and false positives.
🔄 Real-time Predictions
Get instant predictions through our RESTful API with single observation inputs.
📈 Robust Preprocessing
Automated data cleaning, scaling, and handling of missing values for reliable results.
💾 Export Results
Download your predictions as CSV files for further analysis and documentation.
🛠️ Technology Stack
Our project is built using cutting-edge technologies and frameworks:
📊 Model Performance
🔬 How It Works
1. Data Collection
We source observational data from NASA's Kepler and TESS missions, including transit parameters, stellar properties, and planetary characteristics.
2. Preprocessing
Our system automatically handles missing values, converts categorical variables, removes outliers, and applies robust scaling to prepare data for analysis.
3. Machine Learning Classification
Random Forest classifiers analyze the preprocessed data to classify observations into three categories:
- CONFIRMED - Validated exoplanets
- CANDIDATE - Potential exoplanets requiring further study
- FALSE POSITIVE - Signals that are not exoplanets
4. Results & Export
Users receive instant predictions with classification labels and can export results for documentation and further analysis.
🏆 NASA Space Apps Challenge 2025
This project was developed for the NASA Space Apps Challenge 2025, an international hackathon where teams worldwide collaborate to solve challenges using NASA's open data.
Our goal is to make exoplanet detection more accessible and efficient, empowering researchers, educators, and space enthusiasts to explore the cosmos and discover worlds beyond our own.
👥 Team: Deep Sky Divers
We are a passionate team of developers, data scientists, and space enthusiasts dedicated to advancing exoplanet research through innovative technology solutions.
Our interdisciplinary approach combines expertise in machine learning, astronomy, and software engineering to create tools that push the boundaries of space exploration.