What does it do or how does it work?
The project is an artificial intelligence (AI)–based system that automates exoplanet detection using data from NASA’s Kepler mission. Users can upload CSV files or manually enter parameters, and the machine learning model (XGBoost) classifies each object as a confirmed exoplanet, candidate, or false positive. Then, an LLM model (GPT-4) generates natural language explanations of the results, making the analysis accessible and easy to understand.

What benefits does it offer?
We offer increased speed and accuracy in exoplanet classification, enabling the processing of objects within minutes and democratizing astronomical research. It facilitates the participation of both professional scientists and citizen researchers, promoting learning and collaboration.

What is the expected impact of the project?
We expect it to accelerate the discovery of exoplanets, promote open science, and contribute to public education and participation in astronomy. In addition, it aims to reduce the technical barriers to accessing advanced research tools.

What tools, programming languages, hardware, or software were used to develop your project?
The tools we used:

Languages: Python and React
Frameworks and libraries: XGBoost, Pandas, NumPy, Scikit-learn, FastAPI
AI/LLM: OpenAI GPT-4
Development tools: Git/GitHub, VS Code
Data: Kepler Exoplanet Archive (NASA)

How creative is your project?
We consider it a creative project because it integrates machine learning with an LLM to provide scientific analyses explained in natural language, combining advanced technology with scientific communication.

What factors did your team consider?

Accuracy and scientific validity of NASA data
Accessibility for different user profiles
Transparency and traceability in data management
Promotion of collaboration and openness in astronomical knowledge

2025 NASA Space Apps Challenge

A team