How Notability Scores Are Calculated
May 1, 2026 2026-05-07 1:54How Notability Scores Are Calculated
How Notability Scores Are Calculated
A plain-English account of the formula behind the number you see on every profile.
What it is
Every profile in the Notable O’Dubhdas directory carries a single number called the Notability Score. It’s used only to rank profiles within a category and within a status (living or historical) so that the most-recognised figures surface first when you browse. It is not an opinion, it is not editorial, and it is not a measure of significance to the clan — it is a mechanical sum of public, observable signals. Two people with very different careers can land at very similar scores.
The formula
For a living person, the score is a weighted sum of six log-compressed signals:
+ 0.8 · log10(1 + Wikidata sitelinks)
+ 1.0 · log10(1 + Instagram followers)
+ 1.2 · log10(1 + YouTube subscribers)
+ 0.8 · log10(1 + TikTok followers)
+ 0.6 · log10(1 + X (Twitter) followers)
For a historical person (anyone whose status is set to historical), the social-media terms are dropped and the Wikipedia weighting is fixed at 1.5 / 1.0:
+ 1.0 · log10(1 + Wikidata sitelinks)
Family-enterprise entries (e.g. multi-generational businesses where the entity is the family rather than an individual) are not scored at all and are listed separately.
Why a logarithm?
Raw social-media follower counts span six orders of magnitude across the directory — from a few hundred to nearly a million. A linear sum would have a single celebrity drowning out everyone else. The base-10 logarithm compresses each signal so that gaining your first ten followers counts as much as going from 100k to 1m: each successive ten-fold increase contributes the same fixed amount. This puts a respectable working professional and a global household name on the same scale without the latter erasing the former.
The weights, explained
| Signal | Weight | Why |
|---|---|---|
| Wikipedia pageviews | 1.5 | The single best public signal of cross-domain recognition. Pageviews reflect ongoing curiosity rather than a one-time burst of follows. |
| YouTube subscribers | 1.2 | Higher friction than Instagram — subscribing means committing to future content — so each subscriber represents stronger engagement. |
| Instagram followers | 1.0 | The reference platform for personal-brand reach. |
| Wikidata sitelinks | 0.8 | The number of Wikipedia language editions that have an article on this person. A proxy for international reach. |
| TikTok followers | 0.8 | Heavily algorithmic feed; followers are easy to acquire, so weighted slightly below Instagram. |
| X (Twitter) followers | 0.6 | Followers are weakly correlated with reach since the timeline shift; downweighted accordingly. |
The weights live in WordPress settings (odubhda_fame_weights) and can be retuned from the admin without touching code. The defaults reflect the relative reliability of each signal, not an editorial preference between platforms.
Where the numbers come from
None of these numbers are entered by hand. A scheduled job runs over the directory periodically and refreshes each profile’s signals from public sources:
- Wikipedia pageviews — pulled from the Wikimedia REST API for the article URL recorded on the profile (30-day rolling window for living people, 365-day for historical).
- Wikidata sitelinks — counted directly from the Wikidata entity for the QID on the profile.
- Instagram, YouTube, TikTok, X — current public follower counts read from the handle stored on the profile, scraped on the same schedule.
If a signal is missing or fails to fetch, it contributes zero to the score — it does not penalise the profile.
How the score is shown
Every profile displays its score in the sidebar, alongside a timestamp showing when it was last refreshed. Within a category page (e.g. Musicians or Politicians & Public Servants), profiles are ordered by score, highest first. The same is true on surname-variant archives. A few quirks are worth knowing:
- A score of zero means no signals have been collected yet. It does not mean the person is unimportant; it usually means the cron has not yet run on a freshly-published profile.
- Scores can move up or down between refreshes — a Wikipedia article rewrite can change pageview patterns, and follower counts on social platforms can fall as well as rise.
- Two people with similar scores are roughly equivalent on the public-recognition axis. The exact ordering between them isn’t significant.
Editorial-override profiles
A small number of profiles in the directory have no Wikipedia article and qualify by editorial decision — clan-office holders such as the current Taoiseach, for example, who are notable to the clan even where Wikipedia has not yet caught up. These profiles are scored on their social-media signals only, since the Wikipedia and Wikidata terms are zero. A short note on the profile sidebar marks them as editorial overrides.
This is one of the open questions of the system. Six datapoints in the current corpus have both a Wikipedia article and a known Instagram following, and the relationship between the two is roughly — in log space — log10(views) ≈ 0.4 + 0.74 · log10(IG). With a larger corpus, the score for editorial-override profiles could in future be calibrated against this regression so that a no-Wikipedia profile and a Wikipedia-only profile of comparable real-world reach land at comparable scores. For now the system simply scores what it can observe; an override profile with strong socials can still rank highly, as the intent of the directory is to surface notable people, not to gate-keep on which encyclopaedia has noticed them.
What it deliberately does not do
- It does not weight clan rank or office. The Taoiseach does not get a bonus for being Taoiseach.
- It does not weight category. A musician and a scientist are scored on the same axes.
- It does not weight surname variant. An O’Dubhda, an O’Dowd, and a Doody are scored identically.
- It does not include private signals — no email lists, no Spotify streams, no industry rankings, no awards databases. Only what the open web makes visible.
- It does not measure significance to the clan. That is what the rest of this site is for.
The Notability Score is a navigation aid, not a verdict. Its only job is to make sure that when somebody arrives looking for ‘Famous O’Dowds in music’, the people they’re most likely to have heard of show up first.
