|
In all predictions, high
walkability scores are found in downtown Vancouver and the
neighbourhoods of Kitsilano, Commercial Drive, and Marpole. Moreover,
high values occur in the cities of New Westminster, Coquitlam (River
Springs), and North Vancouver. Not really surprisingly, the lowest
walkability scores are found in the North Shore Mountains and in other
sparsely populated areas, for example in the south eastern parts of the
Greater Vancouver Region.
Differences between the indices are
found in the spatial clustering of low and high walkability scores. In
the map based on “Walkability
Index I”, the highly walkable areas are clustered in downtown
Vancouver, some adjacent neighbourhoods, and in East Vancouver. A
smaller cluster is found in Coquitlam and around Metrotown in Burnaby.
The map based on “Walkability
Index III” shows a similar result. Major
differences can be seen in the map based on “Walkability Index II”,
which contains a large number of CTs with high walkability scores. This
is mainly due to the high influence of the land-use mix z-scores.
Basically, the
mapped “Walkability
Index
IV” shows the same general clusters, but it looks a bit
“patchier” than “Walkability Index I” and “Walkability Index III”:
Medium and even low walkability scores are found next to CTs with high
scores. A quantitative analysis shows that many CTs in “Walkability
Index I”, “III” and “IV” are in the same category (lowest and in the
highest quintile), with more than 80% overlapping of CTs in the lowest
category. The overlapping between “Walkability Index II” and “IV” is
lower, but still at about 70%.
In order to get rid of some outliers
that distort the main information provided by the map based on the
“Walkability Index IV”, the results were smoothed out by creating a 50
m x 50m raster and applying the “Majority Filter” -tool in ArcMap.
Final
smooth map
|
Interactions between
walkability index and
socioeconomic factors
A global
linear regression of the socioeconomic
index on the walkability scores shows a negative
correlation between those two factors (y=8.7-0.985x). CTs characterized
by high socioeconomic status are often associated with a low
walkability. However, the correlation is only moderate (r2=0.39).
Especially in the suburban outskirts of the GVRD, the regression
function overestimates the walkability of the CTs, while the high
walkability scores in central Vancouver is not explained by the
relatively low socioeconomic structure in these neighbourhoods.
This negative
correlation is also very obvious in
the selection of CTs that have the lowest and highest 25% walkability
scores and the lowest and highest 25% socioeconomic status scores. While there are only 4 CTs that fall into the
highest categories in both indices, there are 47 CTs that have a low
socioeconomic status but show a high walkability.
When looking at the most prominent factors that
influence the ranking of these CTs in the walkability index, the
urban-suburban differences become again obvious. While the
neighbourhoods with a high walkability show a high dwelling and
intersection density and also a very high land use mixture, the less
walkable neighbourhoods are characterised by very low values in those
categories. Visual impressions of selected neighbourhoods
|
