Optimizing the locations of bike sharing stations
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Master's thesis
Authors
Date
2021
Department
Major/Subject
Mcode
Degree programme
Information and Service Management (ISM)
Language
en
Pages
71
Series
Abstract
Public transport is essential for citizens to move within a city and bike sharing offers an environmentally friendly and a healthy complementary solution to traditional public transport network. Bike sharing has rapidly risen in popularity, however, its successful implementation requires careful consideration for the placement of the bike sharing stations. The cities of Helsinki and Espoo in Finland have successfully launched their bike sharing systems, yet Vantaa is facing unprecedented difficulties in acquiring users for its bike sharing system. As one of the biggest factors in the success of bike sharing systems is the location of the stations, it is possible that the stations in Vantaa have been placed suboptimally. This thesis has gathered success factors of bike sharing systems and factors affecting the location of bike sharing stations from previous literature. Since many of the factors have been considered as common knowledge and sources for them have not been provided, a machine learning decision tree is utilized to determine the important factors in Espoo and Helsinki which have been proven as particularly well-working bike sharing systems, even on a global scale. After confirmation for the success factors, a geographical information system is used to optimize the locations of the bike sharing stations for Helsinki regional area. The machine learning decision tree identified the following factors affecting the location of a bike sharing station: the number of enterprises and jobs, enterprise size, the number of public transport users, the district, the number of residents, age groups of 10-19 and 30-39. The placement of the bike sharing stations in Vantaa is consistent with that of Espoo and Helsinki, therefore it is unlikely to be the underlying issue for the poor utilization of the bike sharing system in Vantaa. Other possible reasons are suggested and left for the confirmation of further research. The results are in line with previous literature and this thesis provides a few additional factors to consider while planning the locations of bike sharing stations as well as offers a way to utilize decision trees in optimizing bike sharing station locations via binary classification.Description
Thesis advisor
Vilkkumaa, EevaKeywords
bike sharing system, location-allocation, public transport, network analysis, decision tree, machine learning, bike sharing stations, optimizing