Fire Chattalytics Project

A new platform has streamlined fire inspections for commercial properties in the City of Chattanooga.

Factors from at least three different databases have been combined for the Chattalytics Fire Inspection Model which predicts and assigns a fire risk score to commercial buildings so that inspectors can prioritize them for inspection.

The fire risk scores are taken from the innovative Chattalytics Model that was created in partnership between the Chattanooga Fire Department and Office of Performance Management and Open Data (OPMOD).

Members of the OPMOD team built the predictive model from scratch and designed the model results prioritization dataset the inspectors will be using.

“Scores are assigned to each property and the database can be actually numbered in order, specific property by specific property, versus lumped into a more general categories of prioritizations,” explained Senior Firefighter Shawn Hays, a fire inspector in the Fire Marshal’s Office.

The data showed that 126 commercial properties in the city had the highest tier fire risk score. Although the buildings that received lower scores also have risk factors, the highest score group needs more urgent attention for code enforcement inspections.

Immediately filtering the list of 126 properties to discern which have been inspected within the last 24 months, the number decreases to 56, meaning that 70 of these properties (56%) have already had recent code inspections.

The new platform has combined the Fire Prevention Bureau’s designation of target properties of which properties are most likely to have the highest risk scores, however, the new model justifies CFD’s predictions with analytics.

The predictive scoring from the added criteria in this new public database provides a high correlation to the existing policies for the prioritization of properties to be inspected.

“It’s reassuring to be made aware that from an objective source outside of our efforts, we are proactively already inspecting more than half of this predictive list,” Hays said.

A way CFD prioritizes needed inspections of “target hazard” properties consists of variables that include: the potential number of people in a building, if there are people sleeping in a building, whether potential assisted evacuation would be needed if there are instances of self-preservation being impaired, whether there are fire protection systems installed, and other potential risks like the storage of combustible and hazardous materials.

For Hays, a new insight that these analytics provided in his inspection district, is that he was not aware of which particular commercial properties are the oldest or which have had the most automatic fire alarm dispatches which was a criteria that is usually difficult to be made aware of with all of the required inspections for new construction, renovations, State licensing, City permits, and reported safety violations from the public.

“With the City OPMOD team’s creation of this model, the overlap of the actual data-driven type of decision-making and the existing policies of the Fire Marshal’s Office prove a high correlation in identifying and ranking which properties to inspect, unifying both time-tested approaches, complementing both of our departments as progressive, forward-thinking and proactive toward better protecting Chattanooga,” Hays added.

You can learn more about this Chattalytics project on it's project page.