Durham Fire Compass

> Community

What factors explain differences in demand based on population and area patterns?

Overview

Correlates incident density with demographic patterns, land use, and urban activity to explain underlying demand drivers.

Total Population and Incidents

Census tract population distribution across Durham County, showing how many people live in each census tract (a subdivision of a county) using 5-year American Community Survey (ACS) estimates. 2024 is the latest available official estimate.

Some census tracts are special-use tracts with no residential population. They are often used for things such as airports, industrial land, institutional land, etc.

Population vs Incident Rate per 1000 Residents

A negative correlation means higher population areas tend to have slightly lower incident rates. The R² value shows the relationship is weak but noticeable.

Why the relationship looks weak or slightly negative

Key statistical + real-world reasons behind the scatter pattern

Land use
High-population tracts are usually residential neighborhoods, while lower-population tracts often include commercial or industrial areas with fewer residents.
Risk type
Incidents happen where activity is happening, such as shops, factories, and busy roads, not just where people sleep at night.
Time mismatch
Population is from 2024 ACS, while incidents are from 2025–2026, so the datasets don’t perfectly align in time.
Why rate/1000
Using incident rate per 1,000 residents removes bias from large population tracts and makes comparison fairer.
R² meaning
Population is not a strong driver of incidents because it ignores where daytime work, travel, and commercial activity occur, which are what actually generate most calls.

Land Use and Incidents

Number of residents in an area is not a reliable indicator of incidents, since land use also matters. This section compares incident rates across zoning types to see which places generate more incidents.

The zoning data is the most recent available (2026) and is maintained by the Durham City/County Planning Department.

Land Use, Land Area (Acres), and Incident Rate per 100 Acres

Residential zones generally have lower incidents per 100 acres as seen on the scatter plot. This may be because they have fewer commercial activities and lower overall public traffic compared to other zones.

Incident Rate per 100 Acres

Raw incident counts are biased by zone size. This shows how intense incidents are within each land area so zones can be compared fairly.

Log(Total Acres)

Durham has very small and very large zones. Logging shrinks the scale so everything fits on the same chart without large zones dominating.

Negative correlation

Bigger zones don’t always mean more incidents per acre. Many large areas include less active or undeveloped land.

R² (low value)

About half of the pattern is explained by land size. The rest comes from land use type and local activity differences.