Data:
The majority of spatial data was collected from the York Region Open Data catalogue (Open Data York Region (arcgis.com)). The point location of grocery stores in York Region were collected using https://overpass-turbo.eu/. The datasets used are summarized in the following table according to their area of interest.
| Economic Vitality | Profile of House by Dissemination Area
|
| Healthy Communities | Year to Date Community Safety Data
Profile of House by Dissemination Area
|
| Sustainable Environment | Wooded Areas in York Region
Bike Paths in York Region Grocery Stores |
Methodology:
In order to get a better sense of the spatial distribution of the community indicators within each focus area, we decided to do three separate Multi-Criteria Evaluations (MCE’s). This allowed us to separately look at and assess the distribution of amenities across York Region before looking at how they interact all together. This also prevents areas of interest (areas containing a low amount of indicators) from being lost amongst all of the separate criteria.
Economic Vitality MCE:
Housing data was obtained from the York Region Open Data portal at the level of Dissemination Areas for 2016, the most recent dataset. Additionally, low income data was obtained through the CHASS Data Centre in the form of percent Low Income Cut Off at the level of DA for 2016.
These datasets are polygon shapefiles that are then converted to rasters using the Polygon to Raster function in ArcGIS. The variables selected are Median Dwelling Value, Median Household Income, and % Low Income Cut Off. Median dwelling value was selected as it was the closest variable to average housing price, as the latter was not available as a variable. The definition of % Low Income Cut Off is “having to devote a much larger share of income than the average family on the necessities of flood, shelter, and clothing”. It is chosen over other measures of poverty due to its considerations for family size and community sizes. Each raster is reclassified into 5 class based on Jenks interval in order to create the most representative measure.
The three rasters are put into the MCE with different weightings. Median Dwelling Value is given a smaller value of 20% due to the fact that this variable reflects the housing price as far as the housing market is concerned, but does not directly reflect the cost of rent or mortgages. Both Median Household Income and % Low Income Cut Off are variables that reflect the economic vitality of each DA, and as such are given 40% weighting each. Dwellings with lower values were given higher suitability values to reflect the affordability of housing. Median Household Income with higher values were given higher suitability values. % Low Income Cut Off with lower values were given higher suitability values to reflect the economic status of each DA.
Healthy Communities MCE:
Similar to the Economic Vitality MCE, housing and community safety data were obtained from the York Region Open Data portal.
The raw data of the population spending less than 30% income on housing was normalized by dividing the raw number of people in that measure by the population per 100k for each dissemination area. The community safety data contained crime data for every category of crime, and as such was far too comprehensive for the scale of this study. We extracted the data for Residential Breaking and Entering, and Assault as primary signifiers of community safety. Both variables were normalized to the population per 100k for each dissemination area as well. Normalization revealed that the BE rate per 100k people is very low, while the assault rate is much higher.
Each polygon was converted to a raster using the Polygon to Raster function, then reclassified based on Jenks method. For the MCE, due to the rates revealed from normalizing the variables, different weightings were applied for each variable. Given the low rates for BE, it was given a 20% weighting, while the other two variables were given 40% each. Both crime variables were given higher suitability values for lower rates of crime, while % spending less than 30% income on housing variable was given higher suitability values for higher percentages.
Sustainable Environment MCE:
Before beginning to work with the datasets specific to the Sustainable Environment focus, we used the Housing Profile dataset to create a clipping layer outline of York Region. This housing dataset was then converted into a raster layer according to the Dissemination Area so that further analysis of the community indicators could be evaluated according to their dissemination area.
The grocery store point dataset was included within healthy communities with the justification that food deserts within Canadian cities is a prominent issue within the health sector (Larson & Gilliland, 2008). In order to get a sense of grocery store accessibility within York Region, each grocery store was placed within the center of a 1000 m buffer. A distance of 1000 m was chosen as it is :commonly used to distinguish if grocery stores are within walking distance” (Larson & Gilliland, 2008). Furthermore, within a “complete community” mindset, being able to choose to walk to get groceries rather than needing to use a car is a good indicator of a “more complete” community.
The bike path data set contained delineations of all types of bike paths within York Region. We chose to ignore the bike paths characterized as ‘paved shoulders’ or ‘shared roadways’ because the emphasis in these cases is still on cars as the main mode of transportation. After eliminating these types of bike paths from the layer we were left with the following types: advisory bike lane, bike lane, protected bike lane, off-road trail, cycle track, and connection. All of these bike lanes required careful planning for the integration of non-automobile transportation. Interestingly, there was an 103% decrease in shape length between the full bike path and the selected bike path layers. This may be an indication of where further development of community infrastructure should be focused. This question would be interesting to explore at a later date. To assess the accessibility of bike paths within York Region a limited buffer of only 500m was placed around the paths. The rationale was based on personal testimonials of both bikers residing in Vancouver, and local residents of York Region who do not bike a lot. The consensus was (within our very limited study) that unless quality bike paths were located very close to home, they were not a valuable amenity and were unlikely to be used as a major form of transport.
Unlike the grocery store and bike path datasets whose relative value was a question of presence or absence within a certain distance, the value of wooded areas was judged to be valuable regardless of distance. However, the respective value associated with the wooded area would decrease as distance increased. To account for this, the Euclidean distance tool was used to calculate the Euclidean distance to the closest wooded area for each cell. This raster layer was then reclassified so that all areas falling within 1 km were applied to a value of 4, 1-2km to 3, and 2-3 km to 2, and any areas beyond a value of 1. In this case, living close or within a wooded area is an indicator of higher community health. It should be noted that the wooded areas dataset was created using 2019 orthophotography, and as such is likely to contain some inaccuracies due to recent developments and the inability of aerial photography to accurately penetrate the canopy.