Tweeting Traffic: Analyzing Twitter for generating real-time city traffic insights and predictions

Abstract

Crowd sourced road traffic management is an open, unexplored problem in data science. With the growth of mobile communications and social media networks, more people are expressing their traffic situations in real-time. We explore how this social media data can be analyzed to generate valuable insights, useful for traffic management and city planning. Our method utilizes background knowledge from structured data repositories for entity extraction from tweets. We proceed to use this spatio-temporal data for traffic incident clustering and prediction. With accuracy and precision measurements providing encouraging results, we build on our methods and present our Continuous Traffic Management Dashboard (CTMD) system: an automated computer system for generating real-time, historic and predictive traffic insights.