Dataframetrainer

A WIP repository to build a data science programming tutor. This repository is capable to run locally and provides a in-browser Python experience. The website on which this should run later is not setup yet, but will follow soon.

Setup

It is assumed that Docker and python3.11 are installed and usable from the command-line. In order to use docker without sudo:

# creates docker group
sudo groupadd docker
# adds current user to docker
sudo usermod -aG docker $USER

Afterwards just execute:

docker compose up

This starts the backend and the database. To start the frontend, use

cd frontend-svelte && npm run dev

Architecture of this repository

CI Pipelines

todo

A/B Testing

To improve the learning experience for the users and to tailer user-specific learning paths, we try A/B testing for the challenges provided to users. This is done in the following way:

• there are specific tables where users are divided into groups
• we provide just one endpoint get_next_challenge which takes the user_id as input and returns the id of the next challenge
• in the backend we implement a strategy per group, this is done in the implementation of get_next_challenge
• in the future we also plan to provide transparency over which strategy is applied to the user