StationNet: An Algorithm for the Extraction and Visualization of Top-n Correlated Bike Stations in Bike Sharing Systems Big Datasets

Authors : Ahmet Şakir DOKUZ

Pages : 275-287

View : 13 | Download : 8

Publication Date : 2021-02-01

Article Type : Research Paper

Abstract :Bike sharing systems insert ignore into journalissuearticles values(BSS); have emerged as an alternative and environmentally friendly transportation tool that provides short-term bike rental to city residents for their close proximity transportation purposes or sports activities. With the emergence and widespread usage of BSS, BSS operators started collecting bike user-related datasets to benefit from it and to increase service quality. Many application areas are present which use BSS big datasets, such as behavioral analyses, urban pattern discovery, and network analysis of bike stations. A bike station network can be defined as a network where bike stations are nodes and the bike trips of users from a station to another station as edges. The extraction of bike station network provides information about which stations are central, which stations have more in- or out-flows, and which regions of the cities are actively used by bike users. However, the extraction of bike station networks is challenging due to the complexity and different characteristics of bike stations, the requirement of new algorithms and new visualization techniques, and the issues related to efficient handling BSS big datasets. In this study, the concept of bike station network extraction in terms of top-n correlated stations is proposed. In particular, the extraction of a bike station network from BSS big datasets are defined and a new algorithm is proposed for extraction of bike station network, and also a new visualization approach that uses common visualization tools is utilized to represent bike station network on a map which would provide more information than a network without a background information. The proposed algorithm and visualization technique are evaluated using one year BSS big dataset. Experimental results show that the proposed algorithm could successfully extract top-n correlated bike station networks and utilized visualization technique is beneficial.
Keywords : top n correlated bike stations, network visualization, network data mining, BSS big data mining