The Content Surgeon: Agency Dashboard

The Content Surgeon produces content for seven doctors across three client entities, and the question that kept coming up was a simple one nobody could answer fast: how is each doctor actually performing? The numbers lived in YouTube Studio tabs, Instagram insights, and a Drive full of exports. The brief was a single screen that pulls all of it together, per doctor, per platform, and then goes one step past the raw numbers to say what they mean.

1. 01

   The data was scattered across YouTube, Instagram, and Google Drive: three surfaces, three formats, no single view. Anyone wanting a read on a doctor had to assemble it by hand every time.

2. 02

   Raw stats aren't the answer. Subs and views per video are inputs; the agency needed interpretation (what's working, what to make more of, what to stop) without anyone spending an afternoon staring at spreadsheets.

3. 03

   It had to refresh on a predictable cadence without a fragile pipeline. Public YouTube data shouldn't need an API key to pull; the parts that do should degrade gracefully when the key isn't set.

4. 04

   Like every internal tool we build, it had to be cheap to run and survive without a CDN: self-hosted vendor libraries so a corporate network or a flaky CDN never breaks the dashboard.

We built it as a single-file React dashboard fed by a Python scan-and-merge pipeline. One script scrapes public YouTube channel stats (no API key needed), a second pulls per-video data via the YouTube Data API, a third merges everything (plus a cached Google Drive scan) into a single data file the dashboard reads. On top of the raw per-doctor, per-platform numbers sits a Claude-as-analyst layer that reads the stats and writes the interpretation: what's performing, what the pattern is, what to do next. Vendor libraries are self-hosted so nothing breaks behind a corporate firewall. Push to GitHub, Vercel redeploys.

Deployed and used inside the agency. The question "how is this doctor doing?" now has a one-screen answer instead of an afternoon of tab-switching. The interpretation layer means the dashboard doesn't just report; it recommends, in plain language. Built to grow: the roadmap layers in YouTube Studio depth, Instagram via Meta Business, and a cross-doctor synthesis pass as the agency's measurement needs mature.