The initial product of NFPA's data effort targets the information needs of first responders
THE NFPA DATA SOLUTIONS CENTER, the first significant piece in NFPA’s new data analytics initiative, is slated to launch by the end of 2016. The center is a map-based online data platform for first responders that will integrate numerous data sources, such as the National Fire Incident Reporting System, relevant local data, census data, and geographic information system information. The goal is to make local data easier to use and understand and to help fire departments make informed decisions about their operations and strategies, said Kathleen Almand, NFPA’s vice president of research.
To use the NFPA data center, a fire department will log into an online portal where users can view and manipulate an assortment of data visually overlaid on maps of their districts. By the end of 2016, users will be able to see their jurisdiction’s past fire incidents plotted on the map and sort the incidents based on ignition source, injuries, fire spread, and other variables. Future enhancements of the data center will allow departments to overlay census and locally gathered data onto the maps to show neighborhood characteristics such as demographics, population densities, and smoke alarm installations, all intended to provide responders with a clearer picture of fire risks and trends happening in their districts.
“Beyond 2016, the center will develop and expand to build more functionality as we get feedback from various stakeholders as to specific data solutions that will help them be more effective and efficient,” Almand said.
The next addition to the data center’s analytics suite will be the Property Inspection Prioritization (PIP) tool, which is slated for release in 2017. The PIP tool will use data analytics to help enforcers map out and prioritize their building inspection schedule based on numerous risk factors.
The NFPA Data Solutions Center—an early version of the tool's dashboard is pictured here—is designed to make local data easier for first responders to understand and use, informing decisions on operations and strategies.
To develop the program, NFPA data scientists surveyed a number of experienced authorities having jurisdiction (AHJs) to determine which risk factors suggest a property should be inspected right away and which factors indicate that an inspection isn’t as pressing. Using that information to weigh numerous variables, a computer model was developed and is now being trained and validated with assistance and expertise from International Fire Marshal Association members. In time, the model will be able to learn on its own as new local inspection data is incorporated. For instance, if a building fails inspection every time, the model may bump up the property and suggest it be inspected more often. If another building never has a problem, the model may lower it on the inspection priority list.
Ultimately, Almand said, the PIP tool will be able to track those patterns and trends and learn from an ever-growing inspection and incident data set to help AHJs make better decisions about their enforcement or inspection programs.