About The Project

Built for NASA Space Apps Challenge 2025

Our Mission

Our mission is to make the discovery of exoplanets more accessible, efficient, and interactive. By leveraging NASA's open-source datasets and cutting-edge AI/ML techniques, we aim to automatically analyze vast amounts of astronomical data to identify new exoplanets with high accuracy.

We strive to empower both researchers and enthusiasts by providing an intuitive web interface where users can explore, upload, and analyze exoplanetary data. Our goal is not only to accelerate scientific discovery but also to foster curiosity and understanding of the universe. Through innovation, transparency, and collaboration, we hope to contribute to humanity's quest to uncover the hidden worlds beyond our solar system.

Technology Stack
Built with modern, cutting-edge technologies

Frontend

Next.js 15
React 19
TypeScript
Tailwind CSS
shadcn/ui
Recharts

Backend & ML

FastAPI
PyTorch
TabTransformer
scikit-learn
Pandas
NumPy

Deployment

Vercel
Railway
NASA Exoplanet Archive
Key Features
  • TabTransformer Architecture: State-of-the-art transformer-based model optimized for tabular astronomical data, combining attention mechanisms with numerical feature processing
  • Real-time Training Progress: Monitor model training with live batch-level updates, metrics visualization, and the ability to safely close the tab while training continues on the server
  • NASA Exoplanet Archive Integration: Train on real data from Kepler Objects of Interest, K2 Mission Candidates, and TESS Objects of Interest with 15,000+ labeled exoplanet observations
  • Timestamped Model Versions: Each trained model is saved with a timestamp, allowing you to track different training sessions and view current model performance metrics
  • Interactive Predictions: Make predictions via manual parameter input or CSV batch upload, with instant classification and confidence scores
Data Sources
Powered by NASA's open data

Kepler Objects of Interest (KOI)

The Kepler mission discovered thousands of exoplanet candidates by observing stellar brightness variations.

K2 Planets and Candidates

K2 extended Kepler's mission, surveying different regions of the sky for transiting exoplanets.

TESS Objects of Interest (TOI)

TESS surveys the entire sky, identifying exoplanet candidates around nearby bright stars.

Astrelis

Exoplanet AI

NASA Space Apps Challenge 2025

© 2025 Astrelis. Built for Space Apps Challenge.