RESPONSE: What is spatial analysis ? Joe Berry

Tue Nov 26 05:14:48 EST 1996

---------- Forwarded Message ----------

From:	Neil Havermale, INTERNET:neilh at farmsoft.com
TO:	mapinfo-l, INTERNET:MAPINFO-L at CSN.NET
DATE:	24/11/96 19:09

RE:	RESPONSE:  What is spatial analysis ?  Joe Berry

Sender: owner-mapinfo-l at csn.net
Message-ID: <32989D55.62EB at farmsoft.com>
Date: Sun, 24 Nov 1996 11:09:09 -0800
From: Neil Havermale <neilh at farmsoft.com>
Organization: Farmers Software Association
MIME-Version: 1.0
To: mapinfo-l at csn.net
Subject: RESPONSE:  What is spatial analysis ?  Joe Berry
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

MAPINFO-L
[Fwd: What exactly is Spatial Analysis ??]=20
Date: Sun, 24 Nov 1996 09:45:33 -0700 (MST)=20
From: ba_sis at FortNet.org (Joseph K. Berry)
  To: mm at farmsoft.com

Neil--  the following is my response to ..."concisely define spatial
analysis, and produce a list of techniques used by GIS that could be
described as being spatial analysis, all on one piece of paper."  Please
post "whence" the question came.  THANKS ...Joe

>
>Maybe you could answer this and post it to mapinfo-l at csn.net
>
>Neil

______________________________

DEFINITION-- Spatial Analysis involves generating new mapped data based o=
n
spatial relationships within (univariate spatial analysis) and among
(multivariate spatial analysis) existing mapped variables.

"Thematic Mapping" and single map "Geo-Query Selections" involve non-spat=
ial
database management which do not constitute spatial analysis, as no new
spatial information is generated by merely "repackaging" and drafting
existing data independent of its spatial contexts.  Operations such as ma=
p
overlay, spatial interpolation, slope, visual connectivity and optimal pa=
ths
constitute spatial analysis as the procedures consider spatial relationsh=
ips
based on contiguity, shape, pattern, coincidence, intersection, proximity
and neighborhoods.

Four functional groups of map analysis operations can be identified:
1)  RECLASSIFICATION operations assign a new value to each map feature
(polygon or grid cell) on a single map based on the feature's position,
initial value, size, shape, or continuity (clumps).
Reclassification based on initial value encompasses all non-spatial analy=
sis
of mapped data (e.g., thematic mapping and geo-query).  Spatial analysis
examples include convexity and Euler indices of individual map feature
shape, and the fractal dimension indicating the relative edge for an enti=
re
landscape.
2)  OVERLAY operations assign new values summarizing the coincidence of m=
ap
features from two or more maps based on a point-by-point, region-wide, or
map-wide basis.
A typical spatial analysis example involves multiple map selection, such =
as
locating all map locations (raster cells or vector intersected parcels) t=
hat
constitute good animal habitat as gently-sloped (slope map),
southerly-oriented (aspect map) , open vegetation (vegetation map) and be=
low
8000 feet (elevation map).  A region-wide overlay example involves
calculating the average slope (slope map) for each forest parcel (vegetat=
ion
map).  A map-wide example includes spatially evaluating a regression
equation relating crop yield to levels of soil nutrients, such as
phosphorous and potassium.  In this instance, each map is a "variable," e=
ach
cell or intersected parcel is a "case," and the values at each location i=
s a
"measurement," as defined in traditional mathematics.
3)  DISTANCE measurement assigns map values based on simple or weighted
connections among map features including distance, proximity, movement, a=
nd
connectivity (optimal paths, line-of-sight, and narrowness).
Spatial analysis examples include the generation of a response travel-tim=
e
map (network analysis for on-road travel, and effective proximity surface
analysis for on/off-road travel), and a visual exposure density surface
indicating the level of exposure for each map location to the entire road
network in a project area.
4)  NEIGHBORHOOD operations assign map values that summarize conditions
within the vicinity of map locations (roving window) based on surface
configuration or statistical summary.
Examples of spatial analysis based on surface configuration examples incl=
ude
the calculation of slope (spatial derivative) and aspect (direction of th=
e
spatial derivative).  Examples of statistical summary by neighborhoods
include spatial interpolation, surface smoothing (neighborhood average),
anomaly detection (deviation of center value from neighborhood average), =
and
localized variation (coefficient of variation).

The classification of analysis operations is discussed in Topic 5, "A
Framework for GIS Modeling," and Topic 7, "Organizing the Map Analysis
Toolbox," in SPATIAL REASONING FOR EFFECTIVE GIS (Berry 1995; GIS World
Books).  The four classes of map analysis operations are discussed in the
book chapter "Cartographic Modeling: The Analytical Capabilities of GIS"
(Berry in GEOGRAPHIC INFORMATION SYSTEMS AND ENVIRONMENTAL MODELING, 1993=
,
Oxford University Press).  A listing and brief description of twenty spat=
ial
analysis operations categorized by the above four classes is part of the
on-line documentation with the "hands-on" exercises in map analysis in th=
e
tMAP -- Tutorial Map Analysis Package (Berry, 1993; GIS World Books).  Th=
e
listing contains a cross-reference to Tomlin=92s categories of Local, Foc=
al
and Zonal spatial analysis operations (Tomlin, 1990: Prentice Hall
Publishers ) and the operations in ARC/INFO GRID, Genacell, and GRASS=20
systems.

Now for your challenge of definition and listing all on one page (disrega=
rd
the above discussion hyperbola and use a small font).  What do you think?
Joe Berry
joeb at cnr.colostate.edu

______________________
FYI

MidNight Mapper
11/24/96
To unsubscribe from this list, send e-mail to majordomo at csn.net and put
"unsubscribe MAPINFO-L" in the message body, or contact bthoen at csn.net