Occupancy analytics is camera-based software that counts how many people are in a space, a zone, or a building at a given moment, without requiring anyone to badge in or carry a sensor. It is a narrower, real-time cousin of people counting: where a simple people counter tallies entries and exits at a door, occupancy analytics keeps a running total for a defined area and can compare it against a set limit, typically refreshing that count continuously rather than on a fixed patrol schedule.
Key takeaways
- Occupancy analytics runs on existing IP cameras, typically 2 MP to 8 MP resolution at 15–30 fps, rather than a separate sensor grid.
- It reports a live count per zone, not per-person identity, and updates continuously instead of on a fixed patrol interval.
- The number it is checked against is usually an occupant load limit set by a locally adopted building or fire code, not a single fixed national figure.
- The International Code Council, formed in 1994, publishes the International Building Code that underlies most U.S. occupant-load requirements.
- Camera-based counting is one of several methods; turnstiles, badge logs, and Wi-Fi/Bluetooth probing are the other common approaches, each with a different refresh pattern.
What Occupancy Analytics Is Measuring Against
The number occupancy analytics is usually being checked against is an occupant load limit, a figure building and fire codes set for a given room or building based on its size and use. In the United States, the International Code Council, the organization formed in 1994 that publishes the International Building Code among its 15 model codes, is the body whose codes underlie most local occupant-load requirements, though the specific limit for any given room comes from the code edition a local jurisdiction has adopted, not from a single national number.
Why the Limit Is Set Locally, Not Nationally
A jurisdiction adopts a specific edition of the International Building Code, sometimes with local amendments, which is why the same room type can carry a different occupant load limit in two different cities even though both are working from the same base model code. A facility operating across multiple jurisdictions cannot assume one occupant-load number applies everywhere it has a location, which is one reason a live, per-site count matters more than a single company-wide policy number.
How Occupancy Analytics Works
From Video Frame to a Live Count
A camera-based occupancy system processes each video frame, detects the people present in a defined zone, and updates a running total, typically within the sub-second latency class common to on-site video analytics, often cited around 250 ms or less per detection cycle on modern edge hardware. That count is then compared continuously against whatever threshold the site has configured, rather than requiring a person to walk the floor and tally heads.
Zones, Not Just a Single Building Total
Most deployments define more than one zone, a casino floor split by pit, a venue split by hall, or a retail floor split by department, since a single building-wide total is rarely the number a fire marshal, an operations manager, or a staffing plan actually needs. Zone-level counting also lets a site flag one crowded area without treating the entire building as over capacity, and a facility with several halls or floors typically runs the same camera infrastructure across all of them rather than a bespoke system per room.
What Limits Count Accuracy in Practice
Accuracy is affected by the same factors that limit any camera-based detection: a camera mounted too low relative to crowd density will lose people behind others in the frame, low light without adequate illumination degrades detection, and a zone boundary drawn poorly, cutting through a doorway or a walkway rather than a clear open area, produces a count that does not match what a person walking the floor would see. Outdoor or semi-outdoor zones, such as a covered entrance plaza, typically call for the same IP66 or IP67 environmental rating used elsewhere in a site’s camera fleet.
| Counting Method | What It Counts | Refresh Pattern | Requires a Badge or Device |
|---|---|---|---|
| Camera-based video analytics | People present in a defined zone | Continuous, sub-second per frame | No |
| Manual headcount | People present at the moment of the count | Periodic, once per patrol or shift | No |
| Turnstile or door counter | Entry and exit events at one point | Real time, per crossing event | No |
| Badge or access-control log | Entry and exit events per credential | Real time, per swipe event | Yes |
| Wi-Fi or Bluetooth device counting | Nearby mobile devices, as a proxy for people | Periodic, based on device probe-request frequency | No, but requires the device to be discoverable |
How Sites Use the Data
Common Applications
- Fire-code and life-safety compliance for assembly spaces, casino floors, and event venues
- Queue and congestion management at entrances, checkpoints, and retail floors
- Space-utilization data for facilities deciding whether to expand, shrink, or reconfigure a room
- Staffing decisions based on which zones are busiest at which hours
From a Single Number to a Historical Pattern
Beyond the live count, a system that logs occupancy continuously builds a historical record over days and weeks, which is what turns a compliance tool into a planning tool. A facility can compare Friday-evening occupancy against a Tuesday-morning baseline across a full quarter, rather than relying on a handful of spot checks, to decide whether a specific zone genuinely needs more staff, a wider walkway, or an additional entrance during peak hours.
What It Does Not Replace
Occupancy analytics reports a count, not an identity, and it does not replace a fire marshal’s inspection or a life-safety plan. It is a continuous, automated version of a headcount that would otherwise be checked manually and only occasionally, which is a difference in frequency and reliability, not a difference in legal authority.
Frequently Asked Questions
Does occupancy analytics identify individual people?
No. Occupancy analytics reports a count of people present in a zone, not the identity of any individual. It is built to answer how many, not who, which is a different function from access control or facial recognition, and the two are typically run as separate agents even on the same camera feed.
How is an occupant load limit actually set?
An occupant load limit is set by the specific edition of the International Building Code, or a comparable model code, that a local jurisdiction has adopted, sometimes with local amendments. Because adoption happens at the jurisdiction level, the same size and type of room can carry a different limit in two different cities, so a multi-site operator should confirm the figure locally rather than assuming one number applies everywhere.
What is the difference between occupancy analytics and people counting?
People counting typically tallies entries and exits at a single door or gate. Occupancy analytics keeps a running total for a defined zone or building and compares it against a threshold, which makes it closer to a live capacity check than a simple traffic count, though the two are often built on the same underlying detection technology.
Can occupancy analytics run on cameras a site already has?
In most cases, yes, provided the existing cameras meet baseline resolution and frame-rate requirements and support a standard interoperability profile such as ONVIF Profile S. This is why occupancy analytics is usually added as a software capability rather than a new hardware installation, which shortens deployment time considerably compared with installing a dedicated sensor grid.
Does occupancy analytics replace a fire marshal’s inspection?
No. It provides continuous, automated data on how many people are present against a set limit, but it does not replace the legal authority or judgment of a fire marshal’s inspection or a facility’s life-safety plan. It is a monitoring tool that supports compliance between inspections, not a substitute for the inspection itself.
Argu’s people counting agent runs this kind of zone-level count on cameras a site already has, for commercial venues and casino floors that need a live occupancy figure rather than an end-of-day estimate, and pairs with the queue management agent for sites that also need to flag congestion at a specific entrance or checkpoint. Contact Argu to scope which zones matter most on your floor plan.
Last updated: September 2026
