Skip to content

Repository files navigation

🌌 Planex

Planex is an AI-powered exoplanet detection platform built during a 30-hour hackathon as part of the NASA Space Apps Challenge 2025.

It leverages NASA's open-source Kepler, K2, and TESS datasets to automatically identify potential exoplanets using machine learning.

🚀 Features

  • 🪐 AI/ML Model: Trained on NASA's Kepler, K2, and TESS datasets to classify exoplanet candidates.
  • 💡 Automatic Classification: Identifies data points as Confirmed Exoplanets, Candidates, or False Positives.
  • 🌐 Interactive Interface: Users can upload new datasets and view predictions in real-time.
  • FastAPI Backend: Efficient inference and preprocessing for large-scale astronomical data.
  • 🛰️ Built in 30 Hours: Developed end-to-end during a 30-hour hackathon event.

🧠 How It Works

  1. Dataset Input: NASA's open-source Kepler, K2, and TESS mission data are used for training and testing.
  2. Preprocessing: The backend cleans, scales, and prepares transit and orbital data.
  3. Model Inference: The trained model predicts whether each observation corresponds to a real exoplanet.
  4. Visualization: The frontend displays prediction confidence levels and categorized results.

🧩 Tech Stack

Frontend:

Backend:

👉 Backend source code: github.com/Fr33b13101/planex

Releases

Packages

Contributors

Languages