Asteroid Classification

understanding vulnerabilities of asteroids

Asteroid Classification

GOALS of the Project

This project focuses on earth's vulnerabilities on asteroids. Every day hundreds even thousands of meteorite hit earth, so analyzing the bigger ones (asteroids) will let us understand the vulnerability towards other spatial objects. The secondary goal of the project is to implement cutting edge mlops to actual real problems.

Technology

1. Machine Learning

2. Data Science

3. MLOps

Data Collection Process :

The data comprises of several entries on several days on NASA JPL, USA. All the data apis are fully functional and can be used for data gathering straight from them. Personally the data is taken from a kaggle dataset , Link : kaggle/nasa-asteroid-classification . The data is actually stored in json format.

Special Note :

Data has been stored using DVC(Data version Control), so the repository package can be used flexibly without adding the data straight in the repo but fetch from any remote source e.g. AWS S3, GDRIVE, etc. For this case, the data has been stored in GDRIVE.

Directory Structure :

The data follows a strict data science project structure.

.
└── root/
    ├── config/
    │   ├── data
    │   └── models
    ├── data/
    │   ├── external
    │   ├── interim
    │   ├── primary
    │   ├── processed
    │   └── raw
    ├── docs
    ├── models
    ├── notebooks
    ├── references
    ├── report/
    │   └── figures
    └── src/
        ├── data
        ├── features
        ├── models
        └── visualization

Installation and Usage :

All the installation and usage techniques are shared in getting_started.md and in commands.md

Approach :

Will update soon :)

Results:

You can visit reports directory where all the runs are stored. Currently, for some privacy issues, the mlflow runs are not shared in here.

Thanks for visiting :D

Do STAR if you find it useful