ROAD TRAFFIC REAL TIME MONITORING THROUGH TWITTER STREAM

Abstract

Worldwide, the cities and the capitals suffer from a common problem –traffic congestion. It takes lot of time and money. The currently used tools for checking traffic are expensive and even have limitations. Since social networking sites, such as twitter, Facebook, etc. have spread in recent years becoming a new kind of real time information channels people share their information, emotions and opinions in terms of short messages. Twitter is one of the most popular social networking sites. The twitter data, i.e. tweets have a benefit over the other social networking sites that it is generally associated with some meta-data and have the 140 character constraint, thus allowing the news oriented data. This paper proposes a system that makes use of tweets to check out for traffic in real time. The system fetches tweet according to several search criteria, process them by using text mining techniques and finally classifies them using the SVM algorithm. The aim is to assign suitable class label to every tweet as related with an activity of traffic events or not. We use support vector machine as a classified model. Finally we can notify the user of presence of traffic by displaying the highlighted map and provide an alternate way.