U.S. Dept. of Transportation Hazmat Incident Anomaly Detector

Partner(s) U.S. Dept. of Transportation
DSWG Teammates Jude Calvillo and Daniel Schweigert
Repo (if different) https://git.io/vPrtb

Project Detail

This ‘anomaly detector’ is one of two apps the DSWG developed as an answer to the DoT’s challenge at 2016’s Bayes Hack, which we ultimately won (for that challenge). This app aims to help DoT execs identify those states that exhibited an anomalous number of hazmat incidents, after accounting for incident seasonality and trend, for a selected month. Thereafter, the user can click one of the anomalous states in the map to get more context in the form of a time-series of incidents for that state (anomalous months highlighted), as well as hazmat-related news from the selected state and month. Whereas our other app (a predictive model + product) is meant to help DoT execs plan/strategize for future hazmat transportation policy, this app is meant to help them tactically respond to real (vs. perceived) problem areas and/or chronicle them.