Goldilocks is a machine learning project that predicts whether an exoplanet is potentially habitable based on its physical and stellar properties.
The project is built with XGBoost and deployed as an interactive Streamlit web app.

• Interactive UI with sliders to input exoplanet properties.
• Prediction output: Determines if a planet is habitable (1) or not habitable (0) with probability confidence.
• Class imbalance handling using scale_pos_weight in XGBoost (habitable planets are rare!).
• Model Insights tab with visualization of decision trees.
• Validation tab including:
  • ROC Curve with AUC score
  • Confusion Matrix

• Dataset: Exoplanet_Dataset_Cleaned_Filtered.csv
• Features include:
  • Orbital period (days)
  • Planet radius (Earth radii)
  • Planet mass (Earth masses)
  • Equilibrium temperature (Kelvin)
  • Star temperature (Kelvin)
  • System distance (parsecs)

Target column:

• pl_hab → 1 if potentially habitable, 0 otherwise

• Python
• Streamlit – interactive frontend
• XGBoost – classification algorithm
• Scikit-learn – model evaluation
• Matplotlib – visualizations

1. Clone the repository:

   git clone https://github.com/AyushSagar16/goldilocks.git
   cd goldilocks