Canonical URL: https://density.io/resources/co2-vs-people-counting-office-utilization

# Can CO₂ sensors measure office occupancy and utilization?

CO₂ readings can support an occupancy estimate, but the estimate needs a model checked against the spaces being measured. Density supplies purpose-built people measurements for workplace decisions: how many people use an area, when a room is occupied, or whether small rooms mainly serve individuals or groups.

The distinction matters if your office already collects CO₂ data. Having an environmental time series does not automatically give you the people counts or occupied-time records needed to compare space use.

## Why concentration is not a universal headcount

People generate CO₂, but the concentration measured in a room depends on more than how many people are present. NIST's [2022 research on indoor carbon dioxide metrics](https://www.nist.gov/publications/development-and-application-indoor-carbon-dioxide-metric-0) describes expected concentration in relation to ventilation, occupant numbers, their CO₂ generation rate and elapsed occupied time.

A [2023 NIST sensitivity analysis](https://www.nist.gov/publications/sensitivity-analysis-co2-concentrations-ventilation-metrics) also identifies room volume and uncertainty in occupant-related inputs. For occupancy estimation, those dependencies mean the same concentration cannot be assigned one universal headcount across different rooms and operating conditions. A change in the concentration does not, by itself, establish a corresponding change in people present.

For a workplace report, that leaves an important question: does the system supply an occupancy estimate with a defined model, or only a concentration reading that someone has relabeled as occupancy?

## CO₂-based occupancy estimation is possible

Researchers have developed models that estimate occupancy from CO₂ measurements. A [2016 paper by Jiang and colleagues](https://arxiv.org/abs/1607.05962) describes a dynamic occupancy estimator and its treatment of noisy measurements. That is evidence that the approach is possible, rather than proof that any installed CO₂ monitor supplies a dependable people count.

If a proposed system estimates occupancy this way, ask what its model needs to know about the space, how those inputs are established, and how the estimates were checked against observed occupancy. Request evidence for the room types and operating conditions relevant to your deployment. Keep estimated counts separate from the underlying concentration readings in the report.

The same buying discipline applies to a people-counting system: define its output and validate it for the spaces and decisions you have selected. A technology label alone does not establish the measurement you will receive.

## Choose the Density output for the planning question

Density's [workplace sensor guidance](https://support.density.io/hc/en-us/articles/37761559048987-Waffle-in-the-Workplace) distinguishes complementary measurements. Open Area provides people counts within larger areas. Entry measures entrances and exits used to calculate occupancy across a covered boundary. Waffle uses camera-free radar in small spaces and reports zero, one, two or three-plus people.

Those Waffle categories support individual-versus-group analysis. A three-plus reading does not distinguish three people from a larger group, so choose a suitable measurement when exact larger counts are required.

Then define the utilization measure. Capacity utilization relates people present to a stated capacity. Occupied-time rate describes occupied time relative to valid observed available time. Neither percentage is a CO₂ concentration; the report needs the relevant occupancy output, space definition and denominator.

Keep environmental information available for its own purpose. For space planning, start with the physical-use question and evaluate the measurement that answers it.

Bring Density your spaces and the decision you need to make. [Book a demo](https://demo.density.io), or email [sales@density.io](mailto:sales@density.io).

Use the [utilization definitions](https://density.io/resources/4-space-utilization-metrics-to-measure-and-optimize) to choose the measure and the [pilot scorecard](https://density.io/resources/occupancy-sensor-pilot-scorecard) to separate observation coverage from measurement agreement.