TraffTrend

Abstract

Traffic management has had difficulty gaining insights about the traffic situation in a city. Here, we classify the data from social media into various cause-effect pairs to mark problems in a locality at a particular time along with its most prominent causes. For this, we classified data into multiple labels such as congestion, accidents, construction etc. using random forest classifier with an accuracy of 82.3%. Using these labels, we find the traffic problems and their probable causes and map it to the location and time of occurrence. Then, this mapping is used to extract useful traffic trends. Also, we show events happening in real time in our dashboard for a particular location so as to keep the common people updated about current traffic situation at various locations.