Winners vs the field chart #10

Why

Winners are now tagged. Comparing them against everyone else on measurable things turns "how do I prepare" into data.

In this repo today

awards in config/hackmeridian-2026.yaml marks first and runner-up projects, exposed as award on every project. There is no finalist list. Some metrics below need other issues first (first commit time, invocations); until they exist, show "n/a".

What to build

• Groups: first, runner-up and field. Add finalist to Place in hackathon.py if a list is published.
• An API route returning the median and p75 for each group of: time from first commit to first deploy, total deploys, distinct wasm (iterations), invocations, commits in window (window_commits), README length, test file count, has demo URL (homepage) and has video URL. README length, test file count and video URL are new fields collected by github/scan.py, which already reads the repo files.
• A reusable GroupComparison component in apps/web/src/components/charts/ taking the groups with their sizes and the metrics with a value per group, rendered as a compact grouped bar per metric.
• A /compare page (pages/compare.tsx) linked from the home page.

Acceptance criteria

• With the winner groups and the field, the page renders every metric with group sizes shown.
• A metric with missing data for a group shows "n/a", never 0.

See CONTRIBUTING.md for conventions.