NFPA Property Inspection Prioritization tool

Coffee/Social starts at 9 a.m

Training starts at 9:30 a.m

PRESENTATION: NFPA Property Inspection Prioritization tool (PIP)

What is PIP?  PIP is an interactive tool designed to help AHJs evaluate and compare the risk level associated with various properties. PIP—in its current format—asks AHJs to identify the occupancy of a property and then rank that property across six different risk factors. PIP then uses some basic data science techniques to come up with a risk score for each property, and the website allows a user to sort their properties based on this score.

How does PIP do what it does? PIP, in essence, tries to replicate the inspection priorities of an experienced AHJ. We created a training set of data based on a few thousand properties and asked a set of AHJs to assess which of those properties were most “at risk” based on the risk factors. From that set of data, we created a data science model that tries to predict whether an AHJ would predict a property as being “at risk.” Those predictions then determine the decide “Priority Level” for each property. A higher priority level means that our model predicted that an AHJ would be more likely to say a property is “at risk.” It’s important to note that PIP is not actually predicting whether a property will have a fire, violate a code, or face some situation. Rather, PIP tries to predict whether an AHJ would judge a property of being “at risk” of some negative event. So far, we are pretty pleased with the accuracy of the model but have numerous ideas on how it can be improved.

You can view the beta website at www.nfpapip.net.

SPEAKER: Joseph Gochal (Division Manager, Data & Analytics | NFPA)  

JGochal@nfpa.org

Joseph Gochal is the Division Manager of Data & Analytics at the National Fire Prevention Association (NFPA) in Quincy, MA, just outside of Boston.  He oversees a team of analysts, data scientists, and subject matter experts developing new insights and tools in the fire and electrical safety communities.  Before joining the NFPA, Joe spent 11 years in the CIA working on data-driven approaches to combat terrorist tactics then was the Chief Data Scientist for a risk consulting company.  Joe attended Rutgers University and New York University.

Business Meeting starts after lunch, but no earlier than 12:30 p.m.