Big Data analysis of Floating Car Data to identify traffic congestion in urban areas

Traffic congestion has increasingly worsened in recent decades, especially in major metropolitan areas with growing populations. Congestion occurs when demand for space exceeds road capacity. These activities, which require individuals to interact, are vital for economic systems and cannot be avoided. In modern society, these demands are sometimes created at the same time, since many individuals travel to school or work and make other deliveries. This thesis will focus on traffic congestion on public highways caused by cars. Developing urban transportation networks requires identifying places prone to traffic congestion. Traffic data collection technologies have advanced in recent years, and real-time traffic information is becoming the norm globally. Floating Car Data (FCD) has become popular in recent years. FCD is timestamped geo-localization and speed data acquired by moving cars via Internet-connected devices such as mobile phones or GPS. Because of the increasing volume of data collected by FCD, it is becoming more important to locate and extract meaningful traffic-related information from the accumulated historical dataset. Congestion patterns are one example of this kind of data. Despite this, the work is challenging due to the size of the collection and the characteristics of the traffic, such as its complexity and dynamics. This thesis uses Data-Driven techniques and a huge FCD dataset to locate, analyse, and infer urban traffic congestion patterns. This will be done by pulling relevant data from a big FCD dataset. The gathered FCD comprises millions of data from 290,185 vehicles, of which major portion of them are personal automobiles and the minor part, fleet vehicles. Most vehicle data is gathered with the frequency of every minute. This data acquisition is done during the the year 2019 . This project compares the relative and absolute traffic congestion of morning and evening peak hours to off-peak hours based on the average speed (in km/h) of cars on their respective route segments. Finally, the traffic congestion pattern is extracted to show how much various metropolitan zones suffer from traffic congestion so future transportation system improvements may be planned appropriately.