Visualizing Nairobi Traffic From Social Media Data

Abstract

Traffic congestion is one of the greatest challenges faced by developing cities in the world. Nairobi, Kenya is not exceptional. Traffic data collection methods such as cameras, radars are too expensive to install and maintain or are unavailable. Citizens have turned to social networks and application such as ma3route to share traffic information. In this paper, we mine traffic information from tweets send to @ma3route handle using the array explode method. Most tweets received on ma3 route correspond to a particular road status. The tweet is parsed and processed. The mentioned location is then paired with the traffic condition information and is visualized using a map application such as Google maps. All this processed tweet data is stored in a database. The traffic status for a specific location for a specified period can be displayed upon query execution. We are able to analyze the traffic condition of a specific location for a 24 hour period and visualize it, enabling us to see traffic patterns such as peak periods. We verified our processed data against on-the-ground checks and crosschecked against other data sources such as @RoadAlertKE and @kenyatraffic and found that the data matched. We conclude that social media data provides an alternative, cheap and real time source of traffic data.