Simon Fraser University, Geography 455, Dr. Nadine Schuurman and Blake Walker
Matthew Bakker, Kelvin Co, Clayton Crawford, James Mack
Team | Proposal | Acknowledgements | Sitemap


Utilized Data | Utilized Software

Utilized Data
      Only two datasets were necessary in order to make our model operational (Table 1). The first was the LHA boundaries which delineate the 14 administrative sub-regions of VIHA. This is the geographic level at which home support scheduling will take place and thus the only appropriate scale for analysis.
      The road network selected for our study was the National Road Network dataset. While this dataset contains fewer road classes then other competing products, those roads that are excluded are resource or recreational in nature and would not be suitable for use by a home support worker. Additionally, the National Road Network dataset contains the information necessary to use a dual range address locator. For application in the patient cluster model, the address locator had to be comprehensive and align with the patient data format used by VIHA. In order to provide for the most realistic results possible we added a travel cost to each class of road found in the road dataset. The travel costs attributed to each road class were created by standardizing Canadian speed limits to a 0-100 scale and then taking the inverse score. For example, a speed limit standardized to 20 would have a travel cost of 80. As such, the speed with which a home support worker can move about the road network was accounted for.

(Click Image to Zoom)

      Due to privacy and confidentiality issues, client’s addresses were not used for this study. Instead, addresses were created randomly. In order to provide a comprehensive result we created 3,000 random points for the extent of Vancouver Island to perform the cluster analysis. These points were constrained along the road network. Of note, some of the points were generated on road segments that did not have suitable addresses for geocoding as they were located on roads without sufficient address information. These points were excluded from final cluster analysis.


Utilized Software
      The following software programs are utilized in our spatial network cluster model: Microsoft Excel 2010; ESRI ArcGIS 10; SANET ArcGIS 10 toolbox extension V4.1; R Statistical Software V2.15.3; S PLUS Statistical Software V8.2.



© 2013 Bakker, Co, Crawford, Mack. All rights reserved.
This website has been optimized for all desktop and mobile devices in the landscape view.