Identification of (near) Real-time Traffic Congestion in the Cities of Australia through Twitter

Transport congestion is an increasing problem especially for larger cities. Typically traffic conditions are monitored in Australia by state and/or federal authorities using expensive electronic devices/sensors on roads or through CCTV cameras. However there is an alternative and far cheaper way to monitor real-time traffic status on roads: through targeted social media analytics. Social networking sites such as Twitter are hugely popular, public and often real-time in nature. A growing number of people post tweets about their lives and feelings every day and everywhere, often with location-based service information included. In this paper, we present an architecture and novel harvesting and analytics approach that exploits this information to identify near real-time transport congestion. Specifically, we present an algorithm for targeted harvesting of tweets solely on the road network using the definitive road network data for Australia. We then implement spatial-temporal clustering algorithms to identify spatio-temporal clusters of tweets on roads to identify potential traffic congestion. We show the scalability of the solution through the use of the large-scale Cloud facilities offered through the National eResearch Collaboration Tools and Resources (NeCTAR -- www.nectar.org.au) Research Cloud.