Predicting Students' Performance in Online Courses using Multiple Data Sources (Codalab challenge)

Description

This repository contains the material for the associated Codalab challenge of "Predicting students' performance in online courses using multiple data sources".

Usage

Clone this repository and make changes as desired. Keep the structure of the starting kit.

The directory FILES contains additional useful files that will be needed when uploading the challenge to Codalab.

Usage:

`cd starting_kit/utilities`

`python make_bundle.py`

There are 2 ways of running the script:

• with sample data found in the starting kit or

• with the real "big" dataset. In this case run:

  python make_bundle.py starting_kit_dir big_data_dir

References

Project repository and documentation
The competition protocol was designed by Isabelle Guyon.
The starting kit was adapted from a Jupyper notebook designed by Balazs Kegl.