Exoproximo

This repository contains a series of machine learning (ML) and artificial intelligence (AI) exercises using space observation data. Each notebook demonstrates different techniques for analyzing, classifying, and deriving insights from data on near-Earth objects (NEOs) and other celestial bodies. Data sources include NASA and related astronomical databases.

Notebooks

1. neo_analysis.ipynb

• Objective: Perform data extraction, preprocessing, normalization, and feature engineering for NEO spectral data.
• Key Functions:
  • Data Loading: Imports spectral and metadata from Marsset and Binzel studies.
  • Data Merging and Cleaning: Merges spectral data with observational metadata, handling timestamps and missing values.
  • Normalization and Feature Extraction:
    • Normalize reflectance data to correct for measurement variations.
    • Calculate spectral slopes and band depths/centers, revealing potential mineral compositions.
  • Clustering and Anomaly Detection
    • Isolation Forest for anomaly detection.
    • Principal Component Analysis and K-Means clustering for grouping.
• Visualizations: Displays correlation heatmaps to explore feature relationships and performs anomaly detection.

2. neo_api_query.ipynb

• Objective: Retrieve and organize NEO data from online astronomical databases.
• Key Features:
  • JPL Horizons: Queries for closest approaches of selected NEOs.
  • PDS Small Bodies Node: Pulls metadata and observational parameters.
  • Data Integration: Combines API-sourced data with existing NEO data to enhance analysis capabilities.

3. keppler_objs_random_forest.ipynb

• Objective: Build and evaluate a random forest classifier for Kepler Objects of Interest (KOI), focused on exoplanet classification.
• Key Components:
  • Data Querying: Retrieves and filters KOI data from NASA’s Exoplanet Archive.
  • Feature Engineering: Selects relevant features, handles missing data, and normalizes inputs.
  • Classification:
    • Initial random forest classification and performance evaluation.
    • Hyperparameter tuning with GridSearch for optimized classification accuracy.
    • Cross-validation to ensure robustness and prevent overfitting.
• Visualizations: Feature importance plots and classification reports to evaluate model performance.

*** From repo:

pipenv --python 3.11
pipenv install ipykernel==6.28.0 python-dotenv==1.0.0
pipenv run python -m ipykernel install --user --name="da_$(basename $(pwd))" --display-name="da_$(basename $(pwd))"

• in notebook, select kernel and python interpreter