Using Geographically Weighted Regression (GWR) to explore spatial variations in the relationship between public transport accessibility and car use : a case study in Lund and Malmö, Sweden
(2017) In Student thesis series INES NGEM01 20171Dept of Physical Geography and Ecosystem Science
- Abstract
- In Sweden the number of cars per person has increased since the mid-20th century. With negative impacts on both health and the environment, private ownership of vehicles represents one of the major challenges in urban transport. To move travelers from privately owned vehicles to public transport has shown to be beneficial in reducing carbon emissions. However, in order to create policies that attract people towards public transport, data of factors influencing transit choice is crucial due to its validation of planning and investments. Previous studies have shown that physical proximity to public transport stations is one of the critical factors when considering transport choice. Consequently, the aim of this thesis is to analyze novel GPS... (More)
- In Sweden the number of cars per person has increased since the mid-20th century. With negative impacts on both health and the environment, private ownership of vehicles represents one of the major challenges in urban transport. To move travelers from privately owned vehicles to public transport has shown to be beneficial in reducing carbon emissions. However, in order to create policies that attract people towards public transport, data of factors influencing transit choice is crucial due to its validation of planning and investments. Previous studies have shown that physical proximity to public transport stations is one of the critical factors when considering transport choice. Consequently, the aim of this thesis is to analyze novel GPS data to investigate the relationship between public transport accessibility and car use in Lund and Malmö, Sweden. By modelling this relationship with the spatial regression model of Geographically Weighted Regression (GWR), regional variations are allowed and investigated. The results in Lund imply a negative association between public transport accessibility and car use, thus suggesting that car use decreases with a higher public transport accessibility. Furthermore, results in Lund indicate that the spatial regression model of GWR is a better fit to the data than the non-spatial regression model of Ordinary Least Squares (OLS). In Malmö, on the other hand, results imply that public transport accessibility does not have significant impact on car use, and suggests that the GWR model is not a better fit to the data than the OLS model. Consequently, the results in Lund and Malmö do not coincide. Nevertheless, in Lund, where model performance is the highest, results imply that car use decreases with a higher public transport accessibility. This study is one of the first to use individual GPS data together with spatial analysis to investigate the relationship between public transport accessibility and car use. Consequently, this study contributes to the literature on the effects of public transport accessibility on car use and on the use of local spatial analyses in accessibility studies. Such knowledge can be utilized in transport planning to reduce car usage. (Less)
- Popular Abstract (Swedish)
- Sedan 1950-talet har antalet bilar per person i Sverige ökat. Med en negativ påverkan på miljön såväl som på människors hälsa, representerar privat ägande av fordon en av de största utmaningarna i transportplanering. Att flytta resenärer från privata fordon till kollektivtrafik har visat sig fördelaktigt beträffande att minska koldioxidutsläppen, dock kräver denna typ av mobilisering politisk styrning. För att skapa effektiv policy, som lockar resenärer till kollektivtrafik, krävs information om faktorer som påverkar färdmedel, vilket sedan kan användas till att validera detaljerad planering och finansiella investeringar. Tidigare studier har visat att tillgänglighet till kollektivtrafik är en av de faktorer som har störst påverkan på... (More)
- Sedan 1950-talet har antalet bilar per person i Sverige ökat. Med en negativ påverkan på miljön såväl som på människors hälsa, representerar privat ägande av fordon en av de största utmaningarna i transportplanering. Att flytta resenärer från privata fordon till kollektivtrafik har visat sig fördelaktigt beträffande att minska koldioxidutsläppen, dock kräver denna typ av mobilisering politisk styrning. För att skapa effektiv policy, som lockar resenärer till kollektivtrafik, krävs information om faktorer som påverkar färdmedel, vilket sedan kan användas till att validera detaljerad planering och finansiella investeringar. Tidigare studier har visat att tillgänglighet till kollektivtrafik är en av de faktorer som har störst påverkan på valet av färdmedel. Därför är syftet med denna studie att analysera individuell GPS data för att undersöka förhållandet mellan tillgänglighet till kollektivtrafik och bilanvändning i Lund och Malmö. Genom att modellera detta förhållande med regressionsmodellen geografisk viktad regression (GWR) möjliggörs och analyseras regionala variationer i relationen. Resultaten i Lund visar att tillgänglighet till kollektivtrafik har en negativ association med bilanvändning, vilket tyder på att bilanvändning minskar med en högre tillgänglighet till kollektivtrafik. I Lund visar även resultaten att den rumsliga modellen GWR är bättre på att modellera data än OLS, som representerar en icke rumslig linjär regressionsmodell. I Malmö har tillgänglighet till kollektivtrafik inte en signifikant relation med bilanvändning, och den rumsliga modellen GWR är inte bättre på att modellera data än den icke rumsliga modellen OLS. Följaktligen så skiljer sig resultaten mellan Lund och Malmö. Men, i Lund där modellen presterar som högst, så tyder resultatet på att bilanvändning minskar med högre tillgänglighet till kollektivtrafik.
Denna studie är en av de första som använder individuell GPS data tillsammans med rumslig analys för att analysera förhållandet mellan tillgänglighet i kollektivtrafik och bilanvändning. Följaktligen bidrar denna studie till litteraturen angående hur bilanvändning påverkas av tillgänglighet till kollektivtrafik och angående användningen av lokala rumsliga modeller i tillgänglighetsanalyser. Denna kunskap kan användas vid transportplanering för att minska bilanvändning. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8919808
- author
- Andersson, Johanna LU
- supervisor
- organization
- course
- NGEM01 20171
- year
- 2017
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- public transport accessibility, Physical Geography and Ecosystem analysis, Geographically Weighted Regression (GWR), travel survey, GPS data, car usage
- publication/series
- Student thesis series INES
- report number
- 428
- language
- English
- id
- 8919808
- date added to LUP
- 2017-06-30 11:55:21
- date last changed
- 2017-06-30 11:55:21
@misc{8919808,
abstract = {{In Sweden the number of cars per person has increased since the mid-20th century. With negative impacts on both health and the environment, private ownership of vehicles represents one of the major challenges in urban transport. To move travelers from privately owned vehicles to public transport has shown to be beneficial in reducing carbon emissions. However, in order to create policies that attract people towards public transport, data of factors influencing transit choice is crucial due to its validation of planning and investments. Previous studies have shown that physical proximity to public transport stations is one of the critical factors when considering transport choice. Consequently, the aim of this thesis is to analyze novel GPS data to investigate the relationship between public transport accessibility and car use in Lund and Malmö, Sweden. By modelling this relationship with the spatial regression model of Geographically Weighted Regression (GWR), regional variations are allowed and investigated. The results in Lund imply a negative association between public transport accessibility and car use, thus suggesting that car use decreases with a higher public transport accessibility. Furthermore, results in Lund indicate that the spatial regression model of GWR is a better fit to the data than the non-spatial regression model of Ordinary Least Squares (OLS). In Malmö, on the other hand, results imply that public transport accessibility does not have significant impact on car use, and suggests that the GWR model is not a better fit to the data than the OLS model. Consequently, the results in Lund and Malmö do not coincide. Nevertheless, in Lund, where model performance is the highest, results imply that car use decreases with a higher public transport accessibility. This study is one of the first to use individual GPS data together with spatial analysis to investigate the relationship between public transport accessibility and car use. Consequently, this study contributes to the literature on the effects of public transport accessibility on car use and on the use of local spatial analyses in accessibility studies. Such knowledge can be utilized in transport planning to reduce car usage.}},
author = {{Andersson, Johanna}},
language = {{eng}},
note = {{Student Paper}},
series = {{Student thesis series INES}},
title = {{Using Geographically Weighted Regression (GWR) to explore spatial variations in the relationship between public transport accessibility and car use : a case study in Lund and Malmö, Sweden}},
year = {{2017}},
}