Engineering · Occupancy analytics

Live occupancy analytics that survive a procurement review

The reason most occupancy analytics never reach a council, a university or a hospital floor is the procurement review. The fix is structural, the architecture has to be the privacy answer, not a policy clause appended to a surveillance camera.

A dark dashboard interface showing live density chart and zone counts
Live anonymous occupancy per zone, the number a procurement reviewer can sign off on

The reason most occupancy analytics never reach a council, a university library or a hospital floor is not the technology. It is the procurement review. A camera on the wall is a question the legal team will not let you answer cheaply, and most vendors do not have a structural answer to it.

The fix is not a longer retention policy. The fix is an architecture where there is no policy load-bearing in the first place.

01 · WHAT "LIVE" ACTUALLY MEANS HEREWhat "live" actually means here

Live occupancy analytics on the Occivar platform is three things stacked together:

  • Per-zone counts, refreshed every second. Each zone is sensed by an edge device that emits an anonymous integer: how many people are in the zone right now. That number is the only thing that leaves the sensor.
  • Density against capacity. Every zone has a capacity calibrated to its fire-egress plan. Density is just the count over that capacity; the platform tracks it against a configurable threshold matrix.
  • Sub-second responsiveness. Threshold crossings, dwell anomalies and zone hand-offs are recognised in the same tick they happen.

What does not exist anywhere on that pipeline: a face, a demographic, a biometric, a frame stored to disk. The model runs on-chip; the raw video does not leave the device. The data on the dashboard is the only data the platform ever has.

02 · WHY THE ARCHITECTURE IS THE PROCUREMENT ANSWERWhy the architecture is the procurement answer

Procurement reviewers are paid to find the thing that could go wrong. The fastest path through that review is not to argue that the thing won't go wrong. It is to be able to point at the architecture and say, precisely: that class of risk does not exist in this system.

A camera that uploads video has an FOI problem, a breach problem, and a community-trust problem, even if no one ever looks at the video. A sensor that does inference on-chip and emits only anonymous counts has none of those problems, because there is nothing in the pipeline to leak.

That distinction is not policy. It is not contractual. It is structural. And structural answers are the answers that clear procurement.

03 · WHAT THE OPERATIONS TEAM GETS BACKWhat the operations team gets back

The platform side of the analytics is what the operations lead lives in:

  • Live density by zone, all day.
  • Hour-by-hour and day-of-week occupancy curves, with the four-week average overlay.
  • Per-zone dwell distributions, exposed for any time window.
  • A status-of-thresholds view, colour-coded against the matrix the safety planner signed off on.

None of those views require knowing who anyone is. All of them are decisions the operations team can act on.

04 · WHERE THIS FITSWhere this fits

Occivar's occupancy analytics is the foundation of every other feature on the platform: alerts, reports, integrations and the Talk-to-your-data assistant all read off the same counts. The privacy answer at the bottom of the stack means none of those features ever need to argue a separate privacy answer at the top.

If your last occupancy conversation stalled in legal review, it is worth seeing the architecture that doesn't.

If your building can't answer one of these questions yet, we should talk.