Panel Paper: Can Artificial Intelligence Smell a Rat? Developing and Testing a Model for Rodent Detection

Friday, November 9, 2018
Madison A - Mezz Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Peter Casey1,2, Kevin Wilson2 and David Yokum2, (1)Office of the Chief Technology Officer, (2)The Lab @ DC

The number of 311 requests for rodent abatement in the District of Columbia has increased substantially over the past two years. Currently, the Department of Health's Rodent Control team mostly inspects locations where rats have been reported by residents. However, there may be places in the city where rat infestations go unreported, increasing the citywide rodent population. In this project, we develop a predictive model that generalizes data coming in through the city's 311 system to identify locations where rodent infestations are likely. We then select locations for Rodent Control to inspect in order to validate the model’s predictions. If successful, this model will help Rodent Control proactively inspect and treat locations where they are likely to find rats, controlling populations before infestations worsen.

Full Paper: