🎟️ Citizen of Nowhere Fan Attention Index: Methodology

What this measures

The Fan Attention Index is a read of how much public attention a sports team draws online, blending Wikipedia readership with Google Trends search interest. It is not a fan count and it is not a survey. Its base layer is a direct tally of page requests to a team's Wikipedia article, from humans, not bots, across as many as 143 language editions for one club; on top of that, most groups add a normalised read of how often people search for the team.

Sources

Team resolution uses the English Wikipedia API to find each team's canonical article and its Wikidata QID. Every team then passes an entity check against Wikidata: its instance-of and sport claims must actually match the team it is supposed to be, so a resolver mistake (a disambiguation page, a same-named unrelated topic) is caught before it enters the index rather than silently inflating a team's numbers. Pageviews come from the Wikimedia Pageviews API (per-article, all-access, agent=user), which counts only human traffic and excludes bots and spiders, summed across every language edition listed on the team's Wikidata item. Search interest comes from Google Trends, normalised within each group so it can be combined with the Wikipedia signal. A Reddit subscriber count is part of the design (a third input, alongside Wikipedia and Trends) but has no live data source wired up yet; every team currently carries a null for it, and the blend below reweights around its absence rather than treating it as a zero.

Window

The published index covers twelve full months, October 2025 to September 2026, at monthly granularity. Every team is measured on the same window, and the baseline below is computed from the same twelve months, so the comparison across teams and leagues is apples to apples.

The board on the index page opens on the latest complete month and can show any single month on record, the last 12 months, or a custom range of 1 to 12 months. A period is scored the same way: its months are summed with the busiest month counted twice (a single month is just that month), in-flux teams keep their 0.5x weight, and each league's cross-sport scale uses the same revenue anchor, pro-rated to the period, which leaves the 0 to 100 scores and ranks unchanged by the pro-rating. Google Trends is a twelve-month pull for one fixed window, so the blend applies to the last 12 months only; every other period is ranked on Wikipedia views alone, and the board says so. The valued-vs-attention multiple is fitted on the twelve-month index and is shown only there.

The baseline

The baseline is the sum of a team's all-language Wikipedia views over the twelve months, plus its single busiest month counted a second time. Every month counts, so a team that is read about all year ranks on that steady attention, while the extra weight on the peak month rewards the moment a team genuinely captured the world's interest (a title run, a trade, a rename) without letting one news cycle swamp the year. Version 0.8 replaced the earlier median-month and quadratic-mean rules with this one.

A custom range of two to twelve months is built the same way: the sum of its months plus its busiest month. A single month is simply that month's views, since its peak is itself and counting it twice would change no score or rank.

The blend: Wikipedia, Trends, and Reddit

Within a group, the design target is Wikipedia at 0.4, Google Trends at 0.3 and Reddit at 0.3. Reddit has no live data yet, so those two-thirds of that weight are not simply dropped: the two working inputs are reweighted proportionally to fill the gap, which comes out to roughly four-sevenths Wikipedia and three-sevenths Trends in practice. That is why the index currently marks every team's signal as either "wiki only" (Trends did not clear the adoption gate below, so Wikipedia stands alone) or "blend" (Wikipedia and Trends combined); there is no "full blend" state yet, since Reddit is not live.

Groups scored on Wikipedia alone

Google Trends measures search volume for a query string, and a team name that collides with an unrelated common word or place name pulls in search volume that has nothing to do with the team. Como, the Serie A club, shares its name with Lake Como; a search-volume signal for "Como" is mostly tourism, not football interest. Rather than build a per-team exception list, the whole Football group is scored on Wikipedia alone, along with AFL, F1, NRL, college football and college basketball, where the same collision problem is common enough (school nicknames, city names, ordinary English words) that Trends would add noise rather than signal. Every team in those groups is a "wiki only" signal, by group policy, not because Trends was individually checked and rejected for that team.

The adoption gate

Even in a group where Trends is used, an individual team's Trends read is only folded into its score once that signal clears a minimum adoption gate: too little search volume to be statistically meaningful is worse than no signal at all, since a thin, noisy series can swing a team's score around for reasons that have nothing to do with its actual attention. A team below the gate falls back to "wiki only" for that one team, even inside a group where most other teams blend successfully.

Spike ratio

The spike ratio is a team's busiest month divided by its median month, over the displayed twelve-month window, computed from the Wikipedia series. A ratio at or above 2.5 is marked with a small lightning icon on the index: it flags a year with one unusually loud month, not a team with weak underlying interest. A high spike ratio can also mean a genuine, sustained shift, such as a rename or relocation, rather than a single-month event. Read the marker as a prompt to look closer, not as a verdict.

In-flux teams: the 0.5x weight

A small number of teams changed name, city or league status recently enough that their Wikipedia history spans more than one identity: a new franchise, a relocation, or a rename. Those teams draw look-up curiosity on top of ordinary fan interest, readers checking "wait, who are the Utah Mammoth" or searching out a freshly relocated club, and that curiosity inflates raw attention in a way that is not really interest in the team as a going concern. To correct for it, an in-flux team's contribution to attention is weighted at 0.5x rather than counted in full. Those teams are marked on the index with a small dot and a tooltip naming the change and the weighting. Read the marker as a flag on that team's number, not as a claim that the underlying interest is fake, only that part of it is name-change curiosity rather than steady-state fandom.

College inclusion rules

College football and college basketball do not have a fixed league roster the way a pro league does, so which programs belong on the index is set by an explicit rule rather than by editorial judgment team by team. Programs are marked with a small double-dagger and a tooltip naming the exact rule when they are on the index for a reason other than plain major-conference membership.

College football: every Power 4 program, Notre Dame, any program with a final AP Top 25 finish in the last 5 seasons, and Army and Navy by name (service-academy programs with a national following that the conference-only rule would otherwise exclude).

College basketball: every major-conference program, plus any program that reached the Final Four in the last 20 seasons or posted a final AP Top 25 finish in the last 5 seasons, on the view that a Final Four run or a ranked season is itself evidence of the national attention this index is trying to measure, even from a program outside the traditional power conferences.

Attention score, and the global scale

The within-group score is the blended (or, for a wiki-only group or team, Wikipedia-only) attention figure, scaled to the top team in its own group: the group leader scores 100.0, and every other team in that group scores its attention as a percentage of the leader's, to one decimal place. That score only compares teams within the same group.

The "Cross-sport score" shown in the All view, and in each category tab when no group chip is selected, instead ranks every team on a single global score. That score is not the within-group attention figure directly; it is that figure scaled by a per-league factor before teams from different sports are put on the same axis. The full method for that scaling, and why it exists, is the next section. The team with the highest cross-sport score anywhere scores 100.0; a team at half that scores 50.0. Most teams score low on this scale, because most teams draw far less scaled attention than the single biggest team in the index; that is the honest picture of how concentrated cross-sport attention is, not a flaw in the scale. The within-group score above, and the within-league or within-group order it produces, is completely unaffected by any of this: the cross-sport scale only changes how teams from different sports compare to each other, never the ranking of teams against others in their own league or group.

Comparing across sports

Raw Wikipedia attention overweights football relative to almost everything else, for reasons that have little to do with fandom. A football club is covered by dozens of language editions of Wikipedia, one per country that follows the sport seriously, while a team in a more regionally-concentrated sport might have five or six. Wikipedia is also football's de facto reference site in a way it is not for US sports, where fans reach first for ESPN, a league's own site or a stats site rather than an encyclopedia article. Compared purely on raw pageviews, a mid-table European club can out-rank a marquee NFL franchise, not because more people care about the club, but because Wikipedia captures a larger share of the club's total attention than it captures of the franchise's.

The fix is a per-league scaling factor, k_L, applied before any cross-sport comparison. For each league L (each Football league counted separately; every other sport grouped by its own group), k_L = the square root of that league's total annual revenue divided by its total Wikipedia attention. A team's cross-sport score is its within-league attention share multiplied by k_L, so leagues whose fans generate more real-world revenue per unit of Wikipedia attention are scaled up relative to leagues (like top-flight European football) whose Wikipedia footprint already runs ahead of their revenue. Revenue figures come from Deloitte's Football Money League (Europe's top 5 leagues plus the WSL), Forbes and Sportico franchise-value reporting, and each league or federation's own published or reputable-aggregator revenue figures elsewhere. Every team carries an anchor_source naming exactly which figure was used and ananchor_confidence of high, medium or low, so a reader can see how solid the revenue side of that team's scaling is.

This scaling only ever touches the cross-sport comparison. It has no effect on the within-league or within-group order: a team's rank among its own league or group rivals is set entirely by its share of that league's or group's Wikipedia (and, where used, Trends) attention, exactly as described above, before k_L is ever applied.

The Football group (every league in it, including MLS and Liga MX) is Wikipedia-only in this version, with Trends left out entirely: club football team names collide constantly with unrelated high-volume searches, for example MLS's Chicago Fire against the unrelated television series of the same name, so no Football league's Trends read can be trusted and none is used.

Valuation, and "Valued vs attention"

Where a team has a franchise valuation in this site's valuations dataset, it is joined to that figure. Every value is converted to USD at the exchange rate on the valuation's own as-of date, not today's rate, so a 2023 valuation in another currency is not silently inflated or deflated by currency movement since then.

Valued vs attention compares a team's valuation with what is typical for its level of attention within its own league. For every team with a valuation, the index first computes a ratio of valuation to attention (the same within-league attention figure its rank in the league is sorted on). It then takes the median of that ratio across every valued team in the league, and divides each team's own ratio by that league median. The result is a multiplier read directly against 1.0, the league norm: ×1.8 means the team is valued at 1.8 times what is typical in its league for its level of attention; ×0.5 means half. A team sitting exactly at its league's typical value-for-attention scores ×1.0. It replaces an earlier fitted-line, percentage-gap version of the same comparison (the fitted line is still computed, but only as information -- see the table below -- and the old residual_pct field is still written from the new multiplier, unchanged in shape).

The multiplier is shown for every league with at least 3 valued teams, by any method including speculative estimates; a league with fewer valued teams than that shows a dash, since a median of one or two ratios is not a meaningful league norm to compare against. This is a change from an earlier version of the index, which additionally required the league's attention-to-value relationship to fit a log-log regression at R squared 0.4 or above -- a bar that left the multiplier blank for entire major leagues (NFL, MLB, NHL, WNBA, F1, both college groups, AFL, NRL) precisely because attention explains value poorly there, which is itself worth showing, not hiding. As of this page's last build:

Shown: AFL, Brasileirão, Bundesliga, Championship, College basketball, College football, Eredivisie, EuroLeague, F1, Handball-Bundesliga, IPL, La Liga, Liga MX, Liga Profesional, Ligue 1, MLB, MLS, NBA, NFL, NHL, NPB, NRL, NWSL, Premier League, Primeira Liga, Scottish Premiership, Serie A, SuperLega, Süper Lig, Top 14, WNBA, WSL.

Shown as a dash (valued teams exist, but fewer than 3 in that league): CFL.

Every valuation names its source under the figure (“Forbes 2026”; tap it for the full source), and CoN estimate marks our own figure. A small badge next to a valuation marks how it was derived: deal means the figure is implied by a recent stake sale rather than a standalone appraisal, athletic dept means the figure covers a university's whole athletic department, not the one team shown, and unofficial (shown muted and italic) means the figure is a third party's estimate, not a reported valuation or a disclosed deal. Estimated figures are excluded from the attention-to-value fit below (they would let the index validate itself against its own guesses), but they DO count toward a league's median ratio and ARE shown a Valued vs attention multiplier, with the unofficial (or CoN estimate) tag repeated next to it, so a reader can weigh it appropriately rather than see a gap in the column. A valuation with none of these three badges is a publisher's own franchise appraisal.

Program values for college football and college basketball (added 2026-09-24): a school's athletic department is valued as one number, but its football and basketball programs earn very different revenue, so showing that one department figure on both rows made them look equally valuable when they usually are not. Where a program-level figure is available (the program value badge), it is built from the U.S. Department of Education's EADA (Equity in Athletics Data Analysis) filing for that school -- the same federal disclosure every Title IX-covered athletic department, public or private, submits every year, broken out by sport. The department's published or estimated valuation is divided by that school's EADA total athletics revenue for the same year to get an implied revenue multiple, and that multiple is then applied separately to the school's EADA football revenue and men's basketball revenue to get each program's own value. A school with no FBS football program (most basketball-only Big East and mid-major schools -- Gonzaga, Villanova, Georgetown and similar) only gets a basketball program value, since there is no football revenue to derive one from. Rows still marked athletic dept have not yet been broken out this way and still carry the whole department's figure; the underlying revenue file, the school-to-EADA-institution match, and the derived program values themselves are in data/fans/college/ and scripts/fans/pending/college_program_values.csv in the repo.

How much attention explains value, by league

Separately from the multiplier above, the index still fits a log-log ordinary least squares line between attention and valuation within each league (non-speculative valued teams only), purely as information about how well attention and value actually track each other -- it no longer decides whether a multiplier is shown. R squared closer to 1 means a league where the fitted line predicts a team's valuation from its attention almost exactly; closer to 0 means attention barely predicts valuation at all, and something else -- a closed, revenue-shared league structure; media-rights deals; ownership scarcity; arena economics -- is doing most of the work instead. Sorted worst fit first, since that end is the more interesting one to read.

Attention explains value well within European football (Premier League, La Liga, Bundesliga, Serie A, Primeira Liga, Brasileirão all sit at R² 0.8 or above), an almost mechanical relationship, likely because European club value is itself driven heavily by global fan reach and media rights, which pageviews also track. It explains value only moderately for the big US pro leagues (NFL 0.26, MLB 0.38) and barely at all for the NHL (0.06) or college basketball (0.00). In revenue-sharing closed leagues like the NFL, franchise value is set largely by a shared national media contract, salary-cap economics and ownership scarcity rather than by any one team's own public following, so value is substantially decoupled from fan attention there -- and that decoupling is itself a finding of this index, not a defect in it.

Limits

• Revenue scaling reflects monetisation as well as fandom. A league's revenue is shaped by media-rights deals, ticket prices, sponsorship markets and other business factors that are not the same thing as how many people care about it, so the k_L cross-sport scale carries some of that business-side distortion along with the fandom signal it is meant to correct for.
• Some revenue anchors are low-confidence estimates. Every team's anchor_confidence says how solid its league's revenue figure is, and several leagues currently sit at medium or low: Argentina's Liga Profesional, Liga MX, the Primeira Liga, Eredivisie and Süper Lig (whose figures are from the 2021-22 season), NPB, the CFL, the Handball-Bundesliga, SuperLega, the college groups, and the IPL (media-rights revenue only, not the franchises' full commercial revenue). A cross-sport comparison involving any of these leagues should be read with that uncertainty in mind.
• Attention is not fandom. Wikipedia and search interest are reference and lookup signals, not ticket-buying or merchandise-buying signals. They capture casual and news-driven interest well and under-represent passionate fan bases that are less likely to look a team up online, particularly younger, social-media-native audiences.
• Attention measures curiosity as well as fandom. A team in the news, for a trade, a scandal, a coaching change, draws look-ups from people who are not really its fans and would not call themselves one. That team can rank above its actual fanbase size on this index simply because more people were curious about it that month. The in-flux weighting above corrects for the most visible version of this (a rename or relocation), but ordinary news-driven curiosity is not otherwise separated out.
• Agreement with a search-based benchmark is not independent proof. This index was checked against an outside popularity benchmark during development, and where the two agree that is treated as supporting evidence for the method, not as proof of it: that benchmark is itself built substantially from search and web-attention signals, so a strong correlation partly reflects shared inputs and blind spots rather than two independent measurements landing on the same answer.
• News and event spikes. Wikipedia and search traffic both react sharply to single events: trades, coaching changes, scandals, deep playoff runs. Even the spike-dampened baseline can still overweight a team that had one unusually loud stretch relative to a team with steadier, lower-amplitude attention.
• Relocations and renames. The Athletics and Utah Mammoth are tracked under their current article titles; both, and every other in-flux team, carry the marker and the 0.5x weighting described above.
• Language skew. US teams skew heavily toward English-language traffic, while European football clubs are followed across dozens of non-English Wikipedias. That difference in language breadth, not just difference in popularity, is a large part of why European clubs lead the index on raw and baseline attention.
• F1 constructor articles. The Mercedes and Aston Martin Formula One articles cover each brand's full history in the sport, not only its current-era entry, since no other dedicated current-team article exists for either. Their attention totals include some volume from earlier F1 stints under the same name.
• Mixed valuation sources and vintages.Valuations come from more than one publisher, with different methodologies for what counts toward a franchise's value, and most but not all figures are from 2026; some entries reflect 2023 through 2025 editions. Cross-league comparisons of the "Valued vs attention" multiplier should be read with that in mind.
• EuroLeague, Top 14, Handball-Bundesliga and SuperLega have no per-club page yet. The site's coverage of these leagues links each club to its metro page, not a dedicated team page, so rows in these four groups do not link out the way other groups do. That is a gap in the site's team pages, not a resolver error.

Updates and history

The index refreshes monthly, on the 3rd. Each month's snapshot is kept, starting from December 2023, so a team's attention, ranking and cross-sport score can be compared month over month rather than only read as a single current figure. That monthly history is browsable directly on the Attention over time page: compare teams, see the biggest risers and fallers within a league, and track whole leagues over time.

The trends page only shows months from June 2025 onward, not the full archive. A bot or scraping campaign had been inflating Wikipedia's recorded "user" pageviews for football pages, and that got corrected in June 2024; since every point on the trends page is a rolling 12-month total, any window ending before June 2025 still has some of that inflated traffic mixed in, so those months are left off rather than shown as if they were clean.

Separately, total human traffic to Wikipedia has been falling over time, mostly because more searches now get answered without a click-through. Left alone, that would make every team and league look like it is losing attention even when its real standing has not changed. The trends page corrects for this by default, dividing each value by that same month's total Wikipedia traffic ("share of Wikipedia attention, per billion views"); a "Raw views" toggle switches back to the unadjusted numbers.

Latest month, and the month so far

The weekly board on the main page ranks each group on the most recent completed month alone: the same all-language Wikipedia views the index uses, but one month rather than the 12-month baseline, so a title run or a star signing shows up in full instead of being averaged down.

Between monthly updates, a provisional "month so far" view refreshes weekly, every Sunday and on the month's last day. It adds up Wikimedia's daily pageview files, which are the same files the official monthly figure is built from, counted the same way. So it is not an estimate of the month: it is the first part of the month's actual count, and on the last day it is the official figure. Ranks move between weekly updates only because the rest of the month happens. Each snapshot also reports how closely the previous month's daily files added up to its official total, as a standing check on that claim.

Citing this dataset

Suggested citation: Citizen of Nowhere Fan Attention Index v0.8 (2026), rankings.citizenofnowhere.org/fans. Released under CC BY 4.0.