Estimating and Differentiating Programming Skills from Unstructured Coding Exercises via Matrix Factorization - Koli Calling 2024

Note: This is a stripped down version of our code prepared for publication along with the paper. It provides the implementation but not the actual data used in our experiments

Requirements

• python 3.11
• pipenv for python virtual env
• setup to run Jupyter Notebooks (e.g. VS-Code + Jupyter Plugins)
• dataset resides in ./data folder

Install packages

• $ pipenv install --dev

Files and Folders

• src/minimal.ipynb: minimal example notebook to load data and train a model
• src/algorithms: implementation of matrix factorization models
• src/shared: code for data loading and preparation
• src/langusage.jar: JAR of Java project developed by our second author, which is used to parse Java code and retrieve language constructs