When the floor was occupied, as events.
A visitor does not arrive one minute at a time. A shop's door counter turns a stream of clicks into "someone came in at ten past nine and left at half past", and this page does the same with the occupied minutes, because a detection rate is counted over events and not over minutes. Each row is one event: when it opened, when it closed, how high it went, how many receivers agreed at the peak, which methods saw it, whether our transmitter's delivered rate collapsed inside it, and which labelled windows it overlaps. Pick one to see the fleet inside it.
The rule. An episode opens on a fleet minute most likely occupied or above, stays open through low, and closes after two consecutive empty minutes or at a gap in scoring. The unit of a detection rate. Method env_inj.
Scope: every day · 2026-09-04 · 2026-08-26 · 2026-08-28 · 2026-08-29 · 2026-08-31 · 2026-09-01 · 2026-09-02 · 2026-09-03 · 2026-09-05 · 2026-09-06 · 2026-09-07 · 2026-09-08 · 2026-09-09 · 2026-09-10
23 episodes on 1 day.
| Episode | Day | From | To | Minutes | Peak (oct) | Agreed | Methods | Rate collapsed | Labels | People | At 90 % |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026-09-04-01 | 2026-09-04 | 2 | 0.91 (to 1.07) | 6 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 96 %, more-than-top 32 %) | ||
| 2026-09-04-02 | 2026-09-04 | 4 | 0.81 (to 1.19) | 6 of 9 | env_inj doppler_inj rate_inj env_pooled |
yes | — | 2 (2 to 3 or more) | 1 or more (mass 100 %, more-than-top 29 %) | ||
| 2026-09-04-03 | 2026-09-04 | 5 | 0.90 (to 1.04) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 97 %, more-than-top 32 %) | ||
| 2026-09-04-04 | 2026-09-04 | 2 | 0.93 (to 1.07) | 6 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 98 %, more-than-top 32 %) | ||
| 2026-09-04-05 | 2026-09-04 | 9 | 0.79 (to 0.90) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3) | 1 or more (mass 100 %, more-than-top 28 %) | ||
| 2026-09-04-06 | 2026-09-04 | 1 | 0.76 (to 0.85) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3) | 1 or more (mass 100 %, more-than-top 28 %) | ||
| 2026-09-04-07 | 2026-09-04 | 2 | 0.77 (to 0.98) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3 or more) | 1 or more (mass 100 %, more-than-top 29 %) | ||
| 2026-09-04-08 | 2026-09-04 | 15 | 0.83 (to 0.93) | 5 of 9 | env_inj doppler_inj rate_inj env_pooled |
yes | — | 2 (2 to 3 or more) | 2 or more (mass 95 %, more-than-top 31 %) | ||
| 2026-09-04-09 | 2026-09-04 | 18 | 0.93 (to 0.99) | 5 of 9 | env_inj doppler_inj rate_inj env_pooled |
yes | — | 3 (3 to 3 or more) | 2 or more (mass 94 %, more-than-top 31 %) | ||
| 2026-09-04-10 | 2026-09-04 | 3 | 0.98 (to 1.06) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 96 %, more-than-top 32 %) | ||
| 2026-09-04-11 | 2026-09-04 | 15 | 0.90 (to 0.96) | 5 of 9 | env_inj doppler_inj rate_inj env_pooled |
yes | — | 2 (2 to 3 or more) | 2 or more (mass 98 %, more-than-top 32 %) | ||
| 2026-09-04-12 | 2026-09-04 | 3 | 0.86 (to 1.03) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3 or more) | 2 or more (mass 91 %, more-than-top 30 %) | ||
| 2026-09-04-13 | 2026-09-04 | 13 | 0.92 (to 1.00) | 5 of 9 | env_inj doppler_inj rate_inj env_pooled |
yes | — | 3 (3 to 3 or more) | 2 or more (mass 98 %, more-than-top 32 %) | ||
| 2026-09-04-14 | 2026-09-04 | 3 | 0.92 (to 1.15) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 92 %, more-than-top 30 %) | ||
| 2026-09-04-15 | 2026-09-04 | 2 | 0.90 (to 1.01) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3 or more) | 2 or more (mass 94 %, more-than-top 31 %) | ||
| 2026-09-04-16 | 2026-09-04 | 2 | 0.88 (to 1.05) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3 or more) | 2 or more (mass 93 %, more-than-top 30 %) | ||
| 2026-09-04-17 | 2026-09-04 | 2 | 0.82 (to 0.86) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3) | 2 or more (mass 90 %, more-than-top 29 %) | ||
| 2026-09-04-18 | 2026-09-04 | 2 | 0.98 (to 1.03) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 97 %, more-than-top 32 %) | ||
| 2026-09-04-19 | 2026-09-04 | 2 | 0.92 (to 1.10) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 92 %, more-than-top 30 %) | ||
| 2026-09-04-20 | 2026-09-04 | 2 | 0.76 (to 0.82) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3) | 1 or more (mass 100 %, more-than-top 27 %) | ||
| 2026-09-04-21 | 2026-09-04 | 2 | 0.89 (to 0.95) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3 or more) | 2 or more (mass 93 %, more-than-top 31 %) | ||
| 2026-09-04-22 | 2026-09-04 | 2 | 0.84 (to 0.92) | 5 of 9 | env_inj doppler_inj |
no | — | 2 (2 to 3) | 2 or more (mass 92 %, more-than-top 30 %) | ||
| 2026-09-04-23 | 2026-09-04 | 2 | 0.99 (to 1.04) | 5 of 9 | env_inj doppler_inj |
no | — | 3 (3 to 3 or more) | 2 or more (mass 97 %, more-than-top 32 %) |
The occupancy app · The instrument behind the count · These episodes as JSON
How to read it
What an episode is
The fleet level per minute is bucketed into empty, low, occupied and high against declared edges. An episode opens on the first minute whose most likely bucket is occupied or high. It stays open through low, because low is the bucket with no labelled minute behind it and is read as "still inside". It closes after two consecutive empty minutes, or at the first unmeasured minute, because a run that crosses a gap would assert something about that gap.
What would refute it
An episode set that does not line up with the researcher's own record of presence on a day he was there, or one that splits a single visit into many fragments because the closing rule is too tight. The labels column is the start of that check: a labelled window with people in it should sit inside one episode, not across three.
What the columns rest on
Peak is the highest fleet level in the run with the top of its receiver-bootstrap band. Agreed counts receivers whose own level read occupied at that minute, so an episode one receiver carried alone reads "1 of 9". Methods lists every level method that reached occupied somewhere inside the run; the delivered-rate collapse is called out because on 2026-09-02 it collapsed inside every occupied episode on four receivers of nine and nowhere else.
Gaps and privacy
An episode is a detector output about a public library floor: a time span and a level, no position, no device, no address. The whole page goes dark while a study is running.