EquiPath

College search tools rank schools by prestige. But for a first-generation student or a student-parent, the question isn’t what’s the highest-ranked school you can get into. It’s where will you actually graduate, without drowning in debt. Federal data makes that concrete: schools serving the highest-income students graduate 74.5% of them in six years; schools serving the lowest-income students, 34.7%.

EquiPath, built with Andre Lee, Davyn Paringkoan, and Leo Zhang for Berkeley’s 7th Annual Datathon for Social Good, scores 4,933 institutions across seven dimensions (affordability, ROI, equity, support, academic fit, environment, and access), personalized to a specific student’s circumstances. A student-parent gets a childcare-adjusted affordability gap. A first-gen student gets amplified support scores at high-support schools. Hard constraints on budget, in-state preference, institution type, and distance filter before any scoring happens.

One rule kept us honest: the algorithm ranks, the AI explains. Rankings come from federal data; Claude translates the numbers into plain language and handles follow-up questions. Worth noting: highly selective schools sometimes surface for low-income students because full-need financial aid makes them among the cheapest options in the dataset. Prestige rankings bury that.

Won first place for Most Creative Project in the Educational Equity Track, sponsored by Snowflake.