Space Apps Challenge 2018

TAG-Galileo | Invent Your Own Challenge

The Challenge | Invent Your Own Challenge

Pose your own challenge, and create a solution of your own choosing!

Our New Home

The project is a web-based application that aids the discovery of habitable planets. Using AI, the app can help scientists confirm if an exoplanet has the ability to sustain human life by predicting the probability of a planet is habitable or not.

TAG-Galileo

Data from NASA Exoplanet Archive and Planetary Habitability Laboratory compiled by Rajeev Misra.

The algorithm is based on 14 properties of the planet and its host star that can determine a planet's habitability:

  • Orbital Period (days)
  • Planet-Star Radius Ratio
  • Fitted Stellar Density (g/cm^3)
  • Planetary Radius (relative to Earth's radii)
  • Orbital Semi-Major Axis (AU)
  • Planet's Equilibrium Temperature (K)
  • Insolation Flux (relative to Earth's Flux)
  • Planet-Star Distance over Star Radius
  • Number of Planets orbiting the Host Star
  • Stellar Effective Temperature (K)
  • Stellar Surface Gravity
  • Stellar Metallicity (dex)
  • Stellar Radius (relative to the Solar radii)
  • Stellar Mass (relative to the Solar mass)

Source: Occurrence and core-envelope structure of 1–4× Earth-size planets around Sun-like stars (2013). "http://www.pnas.org/content/111/35/12655.full"

We used Logistic Regression for the machine learning algorithm because based from our analysis, we need a density estimation-based machine learning algorithm.

Code Repository: https://github.com/JstnClmnt/SpaceApps


References:

[1] https://github.com/rkmisra/cs229_project/tree/mast...

[2] https://exoplanetarchive.ipac.caltech.edu/

[3] http://phl.upr.edu/projects/habitable-exoplanets-c...

[4] http://scikit-learn.org/stable/

[5] https://pandas.pydata.org/

[6] https://matplotlib.org/

NASA Logo

SpaceApps is a NASA incubator innovation program.