Geography Reference
In-Depth Information
Box 44.1
Using GIS to model property valuations
The problem:
what is the value of every property in an
urban area?
Modelling tools and data:
GIS; digitised streets and
small areas; historic rating information; own survey
Principles behind model:
'Tobler's First Law of
Geography' (=neighbourhood effects), plus the effects of
individual property attributes
Model validation:
banded comparison with official
council tax valuation list
Results: (1) high prediction success when compared with
official council tax valuation list (as detailed in Longley
et
al
. 1994); (2) systematic spatial patterning to under- and
over-predictions; (3) GIS provides a potential means for
updating the council tax valuation list.
statistical explanations of housing prices using so-
called 'hedonic' house price models.
This example does not model property prices
from first principles in so far as a central ingredient
to the predictive model was pre-existing rateable
value information. Yet it does provide evidence of
the ways in which GIS provides detailed and
accurate predictions across geographically
extensive areas. The predictions largely replicate
the painstaking survey measurements of a large
number of field surveyors. Perhaps more
important, it also lays the basis for future updating.
House price dynamics are geographically very
variable, reflecting, for example, processes of
obsolescence and decay on the one hand, and
gentrification on the other. Once an explanatory
model has been successfully created for an urban
area, it should prove possible to update the model
predictions to take into account general house
price inflation and within-area variation
attributable to known urban processes. In applied
geography terms, this is thus an instance of a
detailed model of the current form of the city that
might be adapted to incorporate a range of urban
dynamics.
public open space land uses in towns and cities.
This problem lies at the heart of applied
geography and rational planning policy: for
example, to what extent is it reasonable to expect
'brown-field' sites within cities to accommodate
the demands for new house building? Or if
demographic information and projections suggest
that most such households will be small, is it
realistic to expect the demand to be met only by
subdivision of existing properties (e.g. into flats)?
There is much verbiage in the literature about the
quest for more sustainable cities, yet little
comparative work has been carried out on the
ways in which space is filled and quantification of
the effects of population, retail and employment
decentralisation. Indeed, we seem to know very
little about the density of urban living, and the
way that residential areas juxtapose with
employment, leisure and retail land uses. Previous
empirical research clings to vague notions of
apparent declines in the density gradients of cities
over time—which may be no more than an
artefact of the crude ways in which density has
been measured.
In essence, the problem is that there has been
too little detailed work on reconciling the carcass
of the city—that is its built
form
in terms of
housing, industrial, retail, other commercial, and
public open space land uses—with the human
activities that take place in it in terms of journeys
to work, leisure pursuits and the full panoply of
human activities. It follows that the best way of
proceeding is to integrate a data model of the
physical form of an urban area with a data model
that gives some indication of the way in which
human settlement is configured around it. The
most obvious source of information about the
Measuring and simulating urban structure
Our first example provided as an output a spatially
extensive valuation of the urban environment that,
by extension, might be used as an input to
dynamic modelling of relative property price
changes, or even simulation of the spatially
variable effects of changes to local or national
planning policy. A related and established theme
in urban analysis has been the changes in the
density of residential, industrial, commercial and
