Detection of Real-Time Traffic through Twitter Stream Analysis

Social networks have been recently empployed as a source of information for event detection, with specific reference to road traffic activity congestion and accidents or earthquack reporting system. In our paper, we present a real-time detection of traffic from Twitter stream analysis. The system fetches tweets from Twitter as per a several search criteria; process tweets by applying text mining methods; lastly performs the classification of tweets. The aim is to assign suitable class label to every tweet, as related with an activity of traffic event or not. The traffic detection system or framework was utilized for real-time monitoring of several areas of the Italian street network, taking into consideration detection of traffic events just almost in real time, regularly before online traffic news sites. We employed the support vector machine as a classification model, furthermore, we accomplished an accuracy value of 90.75% by tackling a binar classification issue (traffic versus nontraffic tweets). We were also able to discriminate if traffic is caused by an external event or not, by solving a multiclass classification problem and obtaining accuracy value of 80.89%.