Pathways: a Data Driven Course Recommender

Danny Yang, Eric Sun, Sam Fuchs - CDS Insights

Interactive visualization of course progressions through cornell CIS, based on anonymized registrar data. Recommendation system for what courses to take + when to take them.

Update: The website is no longer online or up-to-date, as it was intended as a short-term research project. See this blog post for a writeup about the design and learnings.

Repo Description

Directory:

• unused_vis - unused visualization code for alternative ideas that we explored, retained for reference
• exploratory - files for exploratory data analysis (python notebooks, data extraction script). of these, extract_script.py pre-processses the data, using extract.py to load the original CSV and graph.py to model the graph
• www - source code for pathways webapp (see README in that directory for more info)

Data Preprocessing

• make sure you have python3 and the various packages that our preprocessing uses (nothing fancy)
• cd into exploratory and run python extract_script.py (you can edit the constants at the top of the file to change how the script runs)
• output JSON files will be in exploratory/data, to use them in the production app copy it to www/data

Final Report/Journal

see report.md