Methodology

Every number on this site can be reproduced from public APIs. Here is exactly how.

The formula

Each component is converted to a z-score (its distance from the mean, measured in standard deviations) across the tracked pool before weighting. This keeps the three signals on a comparable scale instead of letting the one with the largest raw numbers dominate.

Component	Weight	Source	Why this weight
Favorite mass	45%	AniList favourites, transformed with log10	Cumulative support is the most stable signal and the hardest to fake at scale.
Wiki momentum	35%	Wikimedia Pageviews, 7-day average vs. prior 30-day average	Captures what is rising now, which a lifetime total cannot show.
Media breadth	20%	AniList popularity of the character's top works	Separates a sustained franchise character from a single-season spike.

Momentum, precisely

For each linked article we take the last 7 days of pageviews and divide by 7 to get a daily average. We then compute the daily average of the preceding 30 days and take the ratio:

Averaging the 7 days avoids a single viral day being read as a trend. The 37-day window is chosen so the two periods do not overlap.

Missing data is not a penalty

If a character has no verified English Wikipedia article, the momentum component is set to neutral (z = 0) rather than a negative value. A coverage gap in our mapping is our limitation, not evidence that a character is declining. This decision is visible in the rankings: characters with large AniList followings but no article can still place highly.

Data sources

• AniList public GraphQL API — character names, images, favorites, and work popularity. No key required.
• Wikimedia Pageviews API — per-article daily pageviews on English Wikipedia. No key required.

Neither source is scraped from a private dashboard, and neither requires a login, so a third party can verify any figure on this site independently.

Known limitations

• Not every character has a Wikipedia article. Coverage is currently 41 of 150. The pool is therefore biased toward long-running and widely documented franchises.
• Name mapping is manual. AniList and Wikipedia spell some names differently, so links are verified by hand. A wrong mapping would silently corrupt a score, so unverified titles are excluded rather than guessed.
• AniList favorites are cumulative. A long-running series accumulates favorites over years, which advantages older characters. The momentum component partially offsets this, but does not eliminate it.
• Wikipedia traffic is not a pure fandom measure. It includes general readers, students, and news-driven visits.
• Updates are periodic, not real-time. A snapshot date is shown at the top of every ranking page.

Reproducing the numbers

The build is a single script: it fetches both APIs, computes the scores, and writes the static pages. The weights are constants in the source, published above on this page.

What this index is not

It is not a paid ranking, not an editorial pick, and not an official popularity chart from any platform. It is a reproducible measurement of public attention, with its method fully disclosed — including the parts where the data is thin.