Authors : Goboitshepo Ororiseng Leroke and Manoj Lall
Abstract: Dealing with traffic congestions is an integral part of a city life. Many hours are spent in traffic congestion leading to great cost and time losses. Normally, traffic conditions are monitored by the government agencies using electronic sensors or CCTV cameras. Undoubtedly maintaining a large networks of sensors and cameras to monitor every street in a city is both impractical and very expensive. However, since, the evolution of social media in all its forms, including blogs, online forums, Facebook and Twitter it is possible to treat social media as a human sensor network. In this study, we present a more cost effective and near real-time traffic monitoring alternative, based on Twitter data analytics which not only reports on the current traffic congestion conditions but also on the reasons causing the traffic congestion. Knowing the cause of the traffic congestion is important as it gives an indication of the severity of the problem. We demonstrate the feasibility of our solution through the use of Twitter data obtained for the city of Pretoria, South Africa. From the data collected, the location and the potential traffic related topics, such as vehicle accident or road construction are extracted. Public sentiments are calculated using a lexicon dictionary based approach and visualized using open street map. This system is aimed at assisting commuters in making an informed decision on route selection.
Goboitshepo Ororiseng Leroke and Manoj Lall, 2019. A (Near) Real-Time Traffic Monitoring System using Social Media Analytics. Journal of Engineering and Applied Sciences, 14: 8055-8060.