Analyzing Terrific Traffic in Urban Areas: A Small Step Towards Bringing Order into City Roads

keywords: Traveling time prediction, crowdsourced data, Google distance matrix API, polynomial regression, scikit-learn
Accurate travel time information enables travellers to plan their journey more wisely and efficiently. This in turn, lessens traffic congestion and improves people's travel experience, particularly in urban areas. Open-source traffic data available from different sources and Google Map API have raised opportunities for analyzing and predicting the traffic more accurately. The purpose of this research work is to analyze bus or car travel time data and showcase different insight and aspect of a society from its traffic pattern. Google Distance Matrix API, Python programming language and machine learning algorithms have been applied in this study to automatically extract, analyze, and visualize traffic data and showcase analysis methodology to improve people's travel experience in Dhaka City and the City of New York. In particular, we apply data analytics to develop an oracle that will give answers to different queries about traffic, such as least congested period and/or least congested route within a day/week/month etc., which in turn would enable people to make informed decisions for travel arrangements. The experimental results and detailed analyses show that there exists a wide fluctuation of travel time during the day in both cities. Furthermore, unlike other works, we accomplish various socio-cultural aspects and behaviour from traffic patterns in those two cities, perform the accessibility analysis and provide recommendations for further research.
reference: Vol. 41, 2022, No. 2, pp. 571–589