Outlier Discovery: find YouTube channels and videos that break out

Outlier Discovery is a three-screen research interface for finding YouTube channels and videos that outperform their own baseline. Surface 01, Discovery, lists channels a discovery agent has surfaced, grouped into five subscriber tiers from under 10k to 400k+ and ranked by relevance, with approve and reject actions. Surface 02, the Leaderboard, ranks tracked channels by a breakout index with a six-point sparkline of its history. Surface 03, Outliers, is a sortable table of videos with their outlier multiple against the channel median, a thumbnail on hover, and a side drawer with channel detail.

The problem and who it is for

The problem it addresses: picking explainer-video topics by watching a feed is slow and biased toward whatever is already big. Comparing a video to its own channel, rather than to the whole platform, surfaces topics that are pulling unusual attention. It is for the researcher or writer choosing what to make next.

Stack and status

Stack observed in the source: React 18 with in-browser JSX, a hand-written CSS file with a light and a dark theme, and view state kept in localStorage. Status: a UI prototype; the discovery agent and the sync pill are represented, not implemented. It is one of the explainer-video tools listed on the homepage.

Live demo

Prototype status: UI demo · every channel, video, view count, breakout index and sync time is mock data generated in the page by a seeded random function

Switch between Discovery, Leaderboard and Outliers in the top bar; click a channel to open its drawer. The demo needs JavaScript.