GIS has come to be essential to urban planning. After
all, we were using overlay analysis when it really was a manual overlay
process consisting of transparent mylar map sheets of various layers of
site information layered on top of a base map rendering. GIS has always
been viewed by the planner as a welcome technological advancement to the
old-fashioned method of site planning. When GIS appeared on the scene it
was positively thrilling to planners to be able to attach data to digital
maps, perform calculations and derive new information to help solve old
problems. Proficiency in GIS, at least in an understanding of its use and
application, is essential to the planning education, and more and more
planning departments are making it a regular and sometimes necessary part
of the curriculum.
GIS, however, has not been very successful at making
spatial analysis widely available to physical and social scientists. My
impression is that most students, researchers, and professional planners
do not have an adequate understanding of what spatial analysis is or what
the issues of spatial analysis are in GIS. Most everyone understands the
graphic nature of GIS and the value of graphical representation and the
information that it can effectively convey (e.g., Tufte). Many of us also
know that the locations of phenomena are important to understanding processes
and many have enthusiastically embraced the computer tools that make thematic
mapping and putting graphic representations in reports easier than could
be done before. These are features that have helped market GIS. However,
most physical and social scientists, let alone thousands of planners in
towns and other government agencies, have no idea how to incorporate spatial
statistics in their analyses or tap into the power that computerized locational
information can add to research.
There are several reasons for this lack of knowledge or expertise as I see it:
a) Geography and the methods of spatial analysis which evolved in that discipline had all but been forgotten after the decline of geography in the United States began in the latter part of this century.
b) Spatial statistics is rarely taught in higher
educational institutions. This may be changing in recent years as colleges
and universities are beginning to catch on to the value of spatial analysis
in research. GIS, as a technological advancement, is probably responsible
for this renewed interest, but there aren’t nearly enough trained professionals
to teach spatial analysis.
c) GIS and spatial analysis still take a lot of
time. This last one is the toughest to overcome. I experienced this as
a recent Ph.D. student.
Incorporating GIS in my dissertation research added many months to my completion time and I ended up spending more time working out techniques and methods of analysis and less on planning theory. Now that I am beginning to mentor Ph.D. students myself and many show enthusiasm for incorporating GIS and spatial analysis in their research, I feel compelled to warn them of the additional effort this will necessitate. Most especially if they’ve never done it before!
GIS and spatial analysis is just not easy to carry out. There is a large learning curve that most physical and social scientists do not want to embark upon especially in the later stages in their careers. Some researchers have created tools and programs (e.g., Anselin, Griffth, etc.) to help make it easier and without them most of us would never be able to accomplish as much as we have. However, widely used GIS software, such as developed by ESRI, is still difficult to learn and does not have the data management and statistical and spatial analytical tools we need to do our jobs. I know this is not news to most of you, but in order to really bring spatial analysis into the classroom, and therefore, into the greater scientific and professional community, we must have better trained analysts and better tools to work with. Additionally, a comprehensive handbook of spatial analysis methods, techniques, and tools, adapted for use in GIS, is needed.
My personal research interests are in urban sustainability and the interaction between population and environment. This demands modeling processes of change which occur through space and time and incorporates data and theories from many disciplines. How to bring together a watershed model, urban development or neighborhood succession models, and the influence of transportation networks in a model of urban change through time is quite a challenge. But the notion of using locational information to unite these different data categories has been inspirational for many urban analysts. After all, all these things occur at some location. Yet there are enormous issues of scale, both temporal and spatial, and data compatibility and accuracy which must be overcome.
What tools are available to model the processes of the social, economic and natural urban environment through space and time? I have recently considered trying data mining software to seek patterns in spatial layers of data from different points in time. Data mining software is available now because of the great interest in business marketing. The software is used to identify consumer patterns and buying behavior of various socio-economic groups. This software is enormously expensive and I can’t say with confidence that the methods presented are applicable for urban change analysis.
That brings me to the last point. Two of the questions asked were whether
spatial analysis is being neglected by the sheer diversity of current research
in GIS and will current research efforts provide an optimum environment
for research in geography, regional science, and other disciplines in the
coming decade? It does appear as if spatial analysis gets lost amidst the
perplexity of seemingly unresolvable issues of data compatibility, scale,
etc. I also think that spatial analysis is hindered by the huge amount
of data that must be handled. However, while recognizing the implications
of these issues on the methods and results of analysis, we must keep moving
toward creating additional data management and analytical tools for the
growing volumes of data we will be accessing in the future. For this reason
more computer science involvement is also needed in developing future GIS.
Indeed, because GIS has applications in such a variety of subject areas
progress will demand multi-disciplinary team efforts.
Telephone: 919-962-4761
Email: ryznar@email.unc.edu