Urban Traffic Analysis Using Social Media Data on the Cloud

Richard O. Sinnott, Yikai Gong, Shiping Chen, Paul Rimba

Abstract: Urban traffic analysis and monitoring is increasingly important with the seemingly, never ending population growth taking place in our cities. Most official work in this area utilizes real traffic data such as vehicle trajectories collected by mobile/fixed sensors on vehicles or at the roadside respectively. However such data are expensive to collect and often unavailable for researchers. In this paper, we explore whether freely available social media data can be used for traffic analysis by comparing the association between official traffic data and social media data patterns on the road network of Melbourne. Given the high volume and high velocity of traffic data, this requires Cloud-based data processing pipelines and associated data models to explore the spatio-temporal relationship between social media data and real traffic data. We identify that social media data can indeed act as a cheaper proxy for more official traffic data sets that are collected by official Government agencies.