A personal research tool for studying which YouTube content formats are gaining audience over time.
What it does
ViralPanel is a small set of command-line scripts, run locally by its single author, that studies
content formats — repeatable title and content patterns such as “how X was built”
or “what X turns into” — rather than individual videos or creators. It maintains a fixed research
panel of roughly 400 public YouTube channels and asks one question: which formats are lifting a
channel’s baseline (its median views), as opposed to producing a single lucky outlier.
How it works
Public metadata and statistics for the research panel are retrieved from the YouTube Data API v3
with an API key only — no sign-in, and no access to anyone's private data.
Separately, the author signs in with OAuth to their own YouTube channel to read its
analytics and to add translated titles and descriptions to their own videos. This authorization
covers that one channel and no other; no other person's account or data is ever accessed.
The tool computes derived analytics locally: per-channel median views and a
“baseline lift” ratio (post-peak median ÷ pre-peak median), used to separate a durable
format from a one-off spike.
Results are stored in a local SQLite database on the author’s own computer and are used
only to inform the author’s own original content decisions.
What it is not
Not a public service: there is no hosted application, no user accounts, and no sign-up.
Not a data product: no YouTube API data is republished, resold, or shared with third parties.
Not an advertising or influencer-matching tool, and it does not upload, modify, or manage
any YouTube content.