Inspiration
Wanting to help students have a greater understanding of their academic future.
What it does
Machine learning powered student risk prediction system that analyzes academic profiles to predict college success with 85% percent accuracy. Our web application provides real time predictions and interactive visualizations to see if they will graduate from college.
How we built it
We used Neo4j with APOC and GDS plugins to extract graph based features from the synthetic UMBC dataset to capture academic network patterns. Our backend combines a ensemble model with engineered features in FastAPI, while our React frontend provides real time predictions and interactive visualizations.
Challenges we ran into
Figuring out how graph databases work and using Neo4j.
Accomplishments that we're proud of
Figuring out how Neo4j worked and what cypher files did.
What we learned
How to implement a ensemble ML pipeline.
What's next for Crash or Pass - Machine Learning College Verdict
Help real students navigate through the difficulties and uncertainties of college life. Advise students on their current college trajectory.
Log in or sign up for Devpost to join the conversation.