Mapping with GIS: When seeing should not always mean believing. Mr Oliver Tomlinson Senior Lecturer in Geographical Sciences School of Education Health and Sciences

1
Mapping with GIS When seeing should not always
mean believing.Mr Oliver TomlinsonSenior
Lecturer in Geographical SciencesSchool of
Education Health and Sciences
2
all maps must tell white lies. And sometimes
these lies are not so little. Maps are
informative, but they can also be deceptive, even
threatening. H.J. de Blij Because of personal
computers and electronic publishing, map users
can now easily lie to themselves - and be unaware
of it. M. Monmonier
3
The problem is no longer limited to those who
have always sought to deceive. GIS and mapping
tools are now in everyday use by those without
GIS, geographic or cartographic training. Even
the experts can get very different answers to
the same questions using different GIS or
different processes within the same GIS.
4
In the beginning What data do we need to solve a
given problem and at what level of
detail? Limitations of humanistic
perception? Issues with common GIS data
sources,...e.g. maps ? Scale of map ?
Cartographic generalisation (e.g. road widths) ?
Hard boundaries (e.g. soil map) ? Currency of
survey V date of publication
5
Data capture Digitising is still a common form of
data capture. ? the process is highly selective ?
further spatial error is added to the data ?
prone to operator bias (e.g. left V right
hander) ? quality standards do exist Existing
digital data ? spatial accuracy non GI
scientists ? attribute accuracy do not
question ? level of generalisation these
issues! fitness for use should always be the
maxim
6
Spatial Representation (e.g. raster data) At
what spatial resolution do we represent the
World? What are potential implications of this
choice?
Real World
Low spatial resolution
Higher spatial resolution
building
lake
river
woodland
7
GIS Analysis I (the black box of wonders) Many
GIS users never question the way the tools they
are using operate and are thus unaware of any
implications. For example, the algorithm
underpinning a given GIS operation may vary
between different GI systems. This means it is
possible to get two different (potentially very
different) answers to the same question using the
same data and methodological approach. This also
applies in a single GIS when the operator is
confronted with options / choices (use defaults?).
8
GIS Analysis II (hard versus soft
approaches) Traditional GIS problem solving
employs hard (yes/no) decision making. But this
has issues. E.g. Find all steep slopes. But when
does a slope become steep? Better approach is
one that uses probability based on things
like fuzzy set membership. Softer decisions, but
more complex!
1.0
0.0
9
GIS Analysis III (seeing is believing?) The
areas used as the basis for study can have a big
impact on the results. Consider poverty. You
could map it at National, Regional, County,
District, Ward or Enumeration District level (and
thats just census areas). Apparent spatial
patterns of poverty can appear and disappear
depending on the areas used. This is termed MAUP
- Modifiable Area Unit Problem. It stems from the
use of arbitrary mapping units for the analysis
of continuous phenomena.
10
Interpreting the results I The results of an
analysis can look black and white.
Source Derbyshire Constabulary, 2003.
Big issues if used to determine practice / policy!
11
Interpreting the results II The GIS display
defaults may not be appropriate and could be
very misleading. An operator may choose display
options which completely hide or over emphasise
spatial patterns. Colour, symbol size, method of
data classification, number of classes, data
aggregation are all factors which are manipulated
in the production of maps. GIS output - Maps are
a form of social control! - they portray specific
social and power relations - they serve their
creators
12
In Summary GIS are very powerful tools which
have become widely accessible. Few users however
have the breadth of expertise required to use
them correctly. Never accept a map produced by a
GIS at face value. Question the data it is based
on, question the way it has been processed and
analysed, question the way it has been
presented.
13
Mapping with GIS When seeing should not always
mean believing.Mr Oliver TomlinsonSenior
Lecturer in Geographical SciencesSchool of
Education Health and Sciences