Analyzing Twitter data is useful for planning and operating transportation systems.
Current University of Toronto research in partnership with Calgary Transit, CUTRIC, SP NA, and NSERC creates a framework for processing a public Twitter feed to identify passenger-related transit “incidents.”
These “incidents” include issues related to the COVID-19 pandemic such as sick passengers, as well as smoking on board vehicles, disruptive passengers, and damaged bus shelters.
Detecting incidents in real time enables transit agencies to immediately respond by dispatching security, safety, or maintenance crews and provide targeted cleaning measures to combat the spread of the COVID-19 virus.
This tweet analysis tool crowdsources incident detection in real time to ensure safe, comfortable, and uninterrupted transit operations.
The research team – MASc student Omar Kabbani, Postdoctoral fellow Willem Klumpenhouwer, and UTTRI associated faculty Professors Amer Shalaby and Tamer El-Diraby – developed a methodology that relies on employing natural language processing to identify eyewitness tweets or in-the-moment tweets about transit.
These tweets are then filtered through a secondary process to identify tweets that pertain to incidents.
This two-step approach provides the flexibility to tailor the tool to detect different issues in the context of transit.