Analysing Commuting Using Local Regression Techniques: Scale, Sensitivity, and Geographical Patterning

Abstract

In this paper, two forms of local regression are employed in the analysis of relations between out-commuting distance and other socioeconomic variables in Northern Ireland. The two regression approaches used are moving window regression (MWR) and geographically weighted regression (GWR). For the first approach different window sizes are applied and changes in results assessed. For the second approach, a Gaussian kernel is used and its bandwidth varied. Seven independent variables are utilised, although a single variable (deprivation) provides the main analytical focus. Differences in results obtained with use of the two approaches are discussed. The relationship between window size or bandwidth size and observed spatial patterning is discussed and elucidated. The results support previous work that indicated severe limitations in using global regressions to examine relationships between socioeconomic variables. Also, the utility of comparing results obtained from MWR and GWR is assessed and the benefits of both approaches are outlined.

Modelling hedonic residential rents for land use and transport simulation while considering spatial effects

“…e GWR approach has recently found use in various applications. ere are examples in the elds of climatology , ecology (Kimsey et al 2008;Zhang and Shi 2004), education (Fotheringham et al 2001), marketing research (Mittal et al 2004), regional science (Huang and Leung 2002), political science (Calvo and Escolar 2003), and transport research (Chow et al 2006;Clark 2007;Hadayeghi et al 2003;Lloyd and Shuttleworth 2005;Nakaya 2001). In the housing eld, there are studies by Bitter et al (2007), Farber and Yeates (2006), Fotheringham et al (2002), Kestens et al (2006), Páez et al (2007), and Yu et al (2007).…”

Geographically weighted regression with a non-Euclidean distance metric: a case study using hedonic house price data

“…However, the choice of distance metric has been largely neglected, 1 and in practice, GWR uses ED to measure 'spatial proximity', although great circle measurements for unprojected geographical coordinates are possible (Charlton et al 2007). Of the few attempts to use a non-ED metric in GWR, Lloyd and Shuttleworth (2005) modified a ward-to-ward distance matrix by calculating ward-to-ward distance with straight-line distances between ward centroids and within-ward distance as the mean distance between enumeration district centroids; Huang et al (2010) introduced a spatio-temporal distance under an ellipsoidal coordinate system for their geographically and temporally weighted regression model.…”

Estimating local car ownership models

“…In this study only a limited geographical data set consisting of 220 wards was used in a GWR model but some interesting insights into local car ownership patterns were apparent. Other Northern Ireland (Lloyd and Shuttleworth, 2005) and a second which attempts to discern patterns in a national Origin-Destination matrix for Japan (Nakaya, 2001). Other transportation studies using GWR include estimating the accident risk at traffic network locations in Toronto Canada (Hadayeghi et al, 2003) …”