Geography Reference
In-Depth Information
Figure 42.4
Market
penetration of Sainsbury.
lives within these buffers is estimated by overlaying
census data on to the spatial backcloth of the
newly created buffers. The problems with this
methodology were discussed in the previous
section. To recap, the difficulties appear when we
try to estimate how much revenue from within
these buffers will actually go to the new store (i.e.
the problem of dealing with in- and outflows from
this buffered region).
The alternative is to build a retail model (as in
equations 42.1-42.2) for Leeds. Demand is
estimated by calculating the spending power
available in each postal district (a function of
population size and affluence). The attractiveness
of the retail outlets is usually given as a simple
function of size. The distance deterrence term is
a function of drive times between each postal
district and each shopping centre or supermarket.
(Birkin and Clarke (1991) offer a simple
introduction to retail models for those new to
the subject.) Having built the model itself, the
next stage is to calibrate it (the procedure to
estimate the model parameters) to reproduce
existing interaction patterns between populations
(either at home or at work) and the grocery
stores of Leeds. The results of this process are
Plate 42.1
Dorothy Perkins and Burton's stores: who
are their targeted customers and where do these
persons live?
Having used geodemographics to locate
customer types and identify possible locations for
new stores, GIS is often used to predict the
revenues of these new stores. Figure 42.5 shows a
classic buffer-and-overlay analysis for a potential
new Kwik Save store in south Leeds. As shown in
Figure 42.5, the GIS buffers a primary, secondary
and tertiary circular catchment area based on
assumptions about how far customers are likely to
travel to the new store. Then the population that
