Modeling Urban Human Mobility and Predicting Planning Transportation Facilities Using K-Means Clustering Algorithm

Abstract

Today, the highest percentage of the world population lives in the cities, and most of the employment opportunities are found in the cities. This brings the challenge to the decision-makers for transportation planning due to the human mobility behavior in the cities. This study aims to analyze urban human mobility patterns through call details records (CDR) data for the development of the K-Means clustering algorithm for planning public transportation facilities in the cities and evaluated the performance of the developed model. This paper used k-means, a machine learning model to cluster the human mobility based on the geographical features from the CDR data and the results of the model showed that the urban human mobility in the city was clustered in three clusters. The results also revealed that there is a need for public transport facilities, especially in the two periods of peak hours identified in the city which is the evening time where it was found that 80 locations of the city, need a high number of public transports means. On the other side of peak hour of the morning which starts, it was found that in 50 locations of the city need a high number of public transports means. The results will enable decision-makers to have an insight into transportation planning in the city for improving the living standards and assuring the quality and sustainability of transport facilities in the city and ensuring speedy exchange of goods and services.