How the Screen Barometer is computed
Data source
All statistics come from CinemaTimes' own showtime aggregation pipeline: we sync the published schedules of cinemas across 11 European countries twice daily and keep a permanent archive of every screening we have ever recorded. The barometer is computed directly from that archive. No box-office, ticketing or admissions data is involved; the barometer measures screen allocation (what cinemas chose to program), not attendance.
Editions and the week window
Each edition covers one completed ISO week: Monday to Sunday, UTC. Editions are computed with the first site build on or after the following Monday and stay stable within the week (the published date on each page is that Monday, not a build timestamp). Statistics begin with the week of 15 June 2026 (our screening archive starts 9 June 2026); week-over-week comparisons begin 29 June 2026.
Distinct-cinema metrics come first
The primary metric everywhere is distinct-cinema reach: the number of different cinemas that played a film at least once during the week. Screen share is that count divided by the market's active cinemas (cinemas with at least one recorded screening that week). Cinema reach is robust: a cinema either played the film or it didn't.
Screening counts are approximate
Total screening counts are honest but approximate. Cinema schedule feeds correct themselves continuously: when a cinema changes a screening's time after we first record it, both the original and the corrected screening can persist in the archive, slightly inflating counts. We label screening totals "approximately" for this reason, and we never rank films by raw screening counts alone.
Climbers, faders and floors
Climbers and faders compare screen share across consecutive weeks, restricted to films that played in both weeks at a floor of at least 5 cinemas in either week (below that, share moves are noise). If the comparison week holds fewer than 500 archived screenings for a market, week-over-week figures are suppressed for that market rather than published against an unreliable baseline.
The Monday epsilon
Editions publish on Monday from data synced through the weekend. Late corrections to the final Sunday of a week can arrive after first publication; subsequent builds within the week absorb them silently. Numbers for the same week can therefore shift by a small epsilon between Monday and the following Sunday. If you cite an edition, the published date shown on each page identifies it.
Citing the barometer
If you cite the barometer, please attribute CinemaTimes and link the country page. Machine-readable editions are planned.