FLAME Moodle Insight Dashboard

FLAME's CDL — manual Excel, every cohort, every term.

FLAME's Center for Digital Learning was tracking pre-orientation engagement by hand: exporting Moodle CSVs, opening them in Excel, calculating engagement scores manually, flagging students for intervention based on instinct. It worked, but it didn't scale, and 'time-to-insight' was measured in hours. The brief was to replace that loop with a tool that gave them the same answers in seconds.

No backend, no API access — by design.

I went local-first as a deliberate MVP strategy. Moodle's API access wasn't immediately available, and even if it had been, every conversation about student data started with FERPA-style privacy concerns. Building it as a client-side app — admins drop a CSV in, parsing happens in-browser, nothing leaves the device — meant we could ship without waiting on IT or risking a privacy review. The CSV parser is a real piece of work: handles arbitrary Moodle export shapes, normalises into a single internal model, and tracks per-row issues for transparency.

Four consultation rounds, weekly stand-ups.

Four sit-down sessions with the CDL team — each one ruthlessly cut features that didn't earn their place, and refined the ones that did. The weighted scoring algorithm and the traffic-light flagging system both came directly out of those conversations: 'we just want to see who's red'. Building tools for actual users in the room is the only way to make sure 'time to insight' goes down instead of up.