From the CID 2018 Challenge video:
The Charlottesville Fire Department regularly assesses fire risks to life and property. With dozens of economic, building stock, demographic, and lifestyle data sets relevant to assessing fire risk available, creating a model to integrate myriad data into a parcel-level assessment will allow dynamic and ongoing monitoring of shifting fire risks in the community. How might we build a model that integrates current and novel data sources into actionable fire risk insights regularly available to local experts in the fire service?
Two locations, Atlanta and Pittsburg, have successfully built and employed data-driven risk assessment models in their daily operations. Other examples that use machine learning to model fire risk are here and here.
The primary challenge can be divided into three parts, each of which are described more fully below.
Predict which structures in Charlottesville are most at risk for a fire event. This will involve joining public data from the Charlottesville Open Data Portal with historical data on fire incidents from the CFD.
CFD assigns each fire incidence a severity level, which is useful for determining an appropriate response. Predict the likely severity level of a fire event for structures around Charlottesville.
How might CFD make information about fire risk available and useful for members of the Charlottesville community? One idea is an interactive website or app that allows people to see the fire risk assigned to their home or place of employment and offers suggestions on what they can do to lower it. What would this look like? What else might be useful?
If you're interested in staying (or getting) involved with this challenge beyond 6/2, please fill out this form with your contact information, level of interest, and area of interest.