Characterizing Bus Travel Time using Advanced Data Visualization Techniques

With the introduction of various automated sensors, traffic data collection has become easier and huge amount of data are getting accumulated over time. One of the interesting challenges in the field of intelligent transportation systems is to effectively utilize such large-scale database. Making meaningful inferences out of this data by conducting in-depth analyses to identify different patterns/trends followed by the traffic variables can lead to the development of more efficient end-applications. The current study analyzes travel time data obtained from buses fitted with global positioning system devices to understand the temporal and spatial variations in travel time in the city of Chennai. For this, data visualization tools such as tree maps and heat maps were used. From temporal analysis, it was observed that travel times are increasing over the years and it was also observed that there is a discernible pattern in travel between weekdays and weekend. From spatial analysis, it was found that there exists a segment specific characteristic of travel time and certain segments experiencing higher travel times in urban areas particularly at intersections. The findings from the study were further used in demonstrating a possible user application, bus travel time prediction system, based on the identified patterns. Performance analysis showed a combination of inputs from same month last year, day of the week, and traffic conditions performing better for the considered dataset.