Fire and safety enforcement officials--or Authorities Having Jurisdiction (AHJs)--face a number of challenges in their code inspection and compliance roles. Limited resources and conflicting priorities often require AHJs to make difficult choices about which properties to prioritize for inspection. The specific factors that go into determining those priorities will likley vary across jurisdictions and merely keeping track of individual properties and risk factors can be a daunting challenge in itself.
To help AHJs with these problems, NFPA developed a decision support platform that we are calling the Property Inspection Priortiziation (PIP) tool. PIP's aim is to assist the prioritization of property inspections by combining assessments of several risk factors with an underlying data science model derived from inputs from over 100 AHJs to "replicate" the prioritization of experienced inspector. PIP, in other words, is an attempt at harnessing the collective wisdom of AHJs to produce a prioritization of properties that would roughly match what an actual AHJ would develop. We recognize that there are many other ways to approach this problem, and PIP is intended to be just one of many tools that an AHJ uses to help set their inspection priorities.
To go into a bit more detail, PIP ranks the inspection priority of a property based on its occupancy type, importance to the community, and other risk factors by employing a common body knowledge shared by AHJs across the US and Canada. The PIP risk factors were formulated with the guidance of the NFPA PIP Development Task Group (which consists of 13 enforcement officials from across the U.S. and Canada). This group identified, in addition to the occupancy type, six high level factors which would have significant impacts on the inspection priority:
Having identified these factors, over 100 expert fire inspectors and AHJ’s then dedicated hundreds of hours in “training” the model that will ultimately drive the ranking system used by the PIP. Right now, PIP can produce prioritization rankings that differ only in one rank from what an actual AHJ would judge. While we find this to be a pretty good initial result, we want PIP to become more accurate over time. Indeed, on of our goal is to enable future versions of PIP to learn from user input to its ability to predict the inspection priorities of properties as perceived by the AHJ’s. You can learn more about PIP on the FAQ page.
We welcome your feedback, questions, and comments about PIP at email@example.com and look forward to hearing from you!