Congestion Correlation And Classification from Twitter and Waze Map Using Artificial Neural Network

Abstract

Traffic congestion has become a big problem in cities around the world, especially in big cities. This causes information about traffic conditions very important to be known by the riders. Such information can be obtained quickly and easily through social media, but not yet known. Previous research has largely focused on classifying congestion data and traffic speed velocity analysis, while the correlation between congestion information from social media and actual traffic flow velocity has not been studied. In this study, we combine data from social media and traffic data collected for 1 week and focus on some major roads in Yogyakarta, Indonesia to investigate the correlation between congestion information in cyberspace through social media and actual traffic speed in Waze applications. The results in this study indicate that the highest precision value of all experiments is 84.01%, while the lowest precision value of all experiments is 0.37%.