Fire has a devastating impact on the neighborhoods of the City of Detroit.

Between January 1st and December 31st, 2015, 2,981 fires broke out in structures across the city. Including fires that spread to other buildings, 2,669 structures were damaged in some way by fire. Residential structures, particularly single family dwellings made up the majority of buildings that burned. Over half of all structures were vacant or had unknown occupancy. Damage severity ranged from minor to collapsed, with 1,328 causing major damage or collapse. 207 additional fires were caused by fires spreading from one building to another. 


Arson was the leading cause of fires in the city, with 53% of fires being suspicious in nature or confirmed arson. 68% of suspicious / arson fires were in vacant structures, and caused a significantly higher level of damage than fires with other causes. 


744 residential structures that caught fire went through foreclosure in 2014 or 2015. Arson rates in vacant foreclosed houses were significantly higher than the city-wide rate, 63% of which were suspicious / arson. 397, or 55% of fires were severe enough to render the house unsalvageable.


Despite the tremendous challenges facing the department, the fire situation in Detroit improved over the course of 2015:

  • The number of structure fires per month peaked over the summer and declined through the rest of the year
  • Response time to fires improved*, limiting the spread of fires to adjacent buildings
  • Suspicious fires decreased through the fall and winter
  • The number of arson investigators increased, resulting in more investigations into suspicious fires and warrants issued*

New leadership at the fire department and additional resources, including demolishing thousands of vacant homes, replacing ten aging fire engines in the busiest parts of the city and reopening closed stations have had the cumulative impact of slowing the fire epidemic, though there remains more work to be done. 


This study consists of three components: data collection, verification, and analysis.


We collect data by recording the digital radio channels used by the Detroit Fire Department to dispatch and direct fire operations with a Uniden BCD536HP digital trunking scanner, which records fire audio to a memory card. A reviewer downloads the audio files and listens to them in sequence on a computer, transcribing the incidents into a spreadsheet with columns for all pertinent data points. The reviewer determines if the fire potentially caused damage to the structure and if so, adds it to the spreadsheet. 

What we do track: Fires that cause tangible damage to a structure that requires repair.

What we don’t track: False alarms, medical calls, minor fires with no damage such as dryer fires or burning food, or garbage fires outside of a structure. We do not track automobile or garage fires unless the fire spreads and causes damage to the primary structure.

One complication is that often a fire in one building spreads to and damages adjacent buildings. if a house fire spreads to the structure next door, how should we count those? Only one fire occurred, but two buildings are damaged.

To fully account for all fires, we use two different terms:

A fire event is a fire that breaks out in a structure and causes damage. Fire events can cause multiple structures to be damaged, but they all share a root cause.

Structure-damaging fires are a subset of fire events. This category covers every building that is damaged by a fire, whether it is an individual fire or caused by adjacent fires (known as exposure or extension fires).

For example:

A fire at 5263 Proctor Street on August 1st, 2015 was severe enough that it spread to two adjacent houses, destroying them, and caused minor damage to four houses directly across the street. If we count just the original fire, we understate the severity of damage caused by the fire. If we count all seven houses as separate fires, it overstates the number of fire set, as seven individual fires did not occur. For this situation we would classify the fire at 5263 Proctor as a fire event, with six additional structure-damaging fires. 

Changes in Methodology

After the release of the midyear fire report, we made some minor changes to the methodology of our survey. The primary change was that we began verifying in person fires that caused little or no damage. In cases where we could not find any damage, these were noted as Damage: None. Some data from the previous report is corrected and updated accordingly. 

Why Loveland's Numbers Won't Match the City's Numbers

The methodology of Loveland's data collection efforts and what we tracked differs from data that is provided by the City of Detroit through its open data portal. There are certain types of fires we did not track, including vehicle and garage fires unless they damaged a primary structure. How Loveland determines fire cause is from the disposition code, which may change after investigation.


As the fire events are added to the spreadsheet, a surveyor is dispatched to the scene of the fire to verify all details and assess the level of damage. The surveyor photographs the fire, talks with neighbors and property owners, and takes notes on all details of the fire that may be of value. This information is collected and sent back to Site Control using the Loveland surveyor app, where it is checked as part of the quality control process. The reviewer then makes a final determination about the level of damage caused by the fire and checks all of the details.

Definitions in Fire data Spreadsheet: 

Date, Time: All audio files collected from the scanner are time stamped. The time is the actual time of dispatch.

Orientation: This field gives details about where the fire was located, which are usually filled in later. EXT is short of extension fire, or a fire spreading from next door. ADD is for additional fires that happen in the same structure in the same month. 2ND denotes a second alarm fire.

Street name, zip code: Geographic information.

PID: Parcel ID. This is the tax assessor ID number for the property, which is how we track exact locations.

Building type: This gives details about what type of building it is, including apartments, single-family dwelling, multi-unit dwellings, commercial buildings, and other specific types.

Occupancy: Occupied, Vacant, Partial Occupancy, Unknown. This is first the stated occupancy of the structure as given by the fire department at the time of the fire. When the property is surveyed, additional information from neighbors or other sources is factored into determining occupancy. If it is unclear if the property was occupied or not at the time of the fire, we use unknown.

Alarm: The number of alarms called for the fire. Mainly for large, out of control fires that require additional companies. We track multiple alarms as fire events so we can note the additional companies being sent.

Damage: Fire damage falls into four different categories.
Minor fire damage is mostly cosmetic in nature, but still requires repairs. This includes smoke / soot damage, light scorching, and water damage. Houses with minor damage are usually inhabitable after the fire. 

Moderate fire damage is significant enough to require immediate repairs, and may temporarily render a house uninhabitable.

Major fire damage is severe enough that it would be costly to repair the structure and makes it uninhabitable. Severe charring inside and outside the building, with damage to contents. Small holes in the roof caused by fire or by cut by firefighters to ventilate the structure.

Collapsed fire damage causes partial or total structural collapse of the roof or walls. This includes buildings that have burned down to the foundation. Walls may still be standing, but parts or all of the roof have caved in.

Disposition: The fire department uses a three-digit code number to assign a cause for the fire. This code is given by the firefighters after the fire is extinguished:

We use disposition codes to determine if a fire was deliberately set. Disposition 011 Incendiary means that evidence of arson was found at the scene. 021 Suspicious means that the fire appears to have been deliberately set. We combine these two numbers to determine a hypothetical suspicious / arson rate.

There are some shortcomings with this method, as it is a judgment call made by fire personnel at the scene based on the information they have at the time. Some fires reported as 000 Unknown / Undetermined are later confirmed to be suspicious by field surveyors talking to neighbors.

Sometimes no disposition code is read over the air; in this case, we note it as “no disposition given.” 

Apparatus designations, including C, E1, E2, E3, E4, L1, L2, S: These are short for battalion chief, first engine called, second engine, third engine, fourth engine, first ladder called, second ladder, and rescue squad. The standard response for a fire is to dispatch a battalion chief, two engine trucks, a ladder truck, and a rescue squad. Depending on the severity of the fire, some companies may be called off before they arrive, or additional companies may be requested. The numbers in the spreadsheet correspond to the company number; C1 is battalion chief 1, E1 is engine company 1, L8 is ladder company 8, and so on. 

With the introduction of mutual aid agreements with the cities of Highland Park and Hamtramck, their fire companies now respond to Detroit fires and vice versa. Highland Park fire companies are denoted as HP1, while Hamtramck companies are HM1. 

Extra: When additional companies are requested above and beyond four engines or two ladder trucks, they are noted in this column. Special units, including senior chiefs, hazardous material, or fire marshal units are noted in this column.

Arson: Denotes if the arson squad is requested.

IEH: Illegal electrical hookups, or people stealing electricity. This can be a contributing factor to occupied residential fires.

BH: Bad hydrants. Some of the city’s fire hydrants are old and no longer work properly. This causes delays in responding to fires, as fire engines carry only carry a limited amount of water in their onboard tanks, and require hydrants to provide water.

Verified: The fire location has been visited by a surveyor, and the data submitted has been checked for accuracy. 

Notes: Reviewers and surveyors compile lengthy notes about the fire based on what can be heard on the digital scanner and what they learn in the field. Some of these notes have been edited to remove sensitive information or personally identifying details.

*Where noted, data was provided by the Detroit Fire Department independently of Loveland research efforts.