Transcript CPSC 601.82 Lecture 8

GIS Software
Dr. M. Gavrilova

What is a GIS system?
A system containing spatially referenced data that can
be analyzed and converted to new information for a
specific purpose.
GIS software:
 ArcInfo
 GeoSQL
 SmallWorld
Statistical analysis software:
 S-Plus Extension for ArcView GIS 3.2
 SpaceStat (TerraSeer)
 Grass GIS
GIS visualization software:
 ArcView, ArcGIS, Visual_Data, GIS Viewer, etc.

Oracle Spatial

PostgresSQL

SDE

Traditional Relational Database

Spatial Processing

What is the difference? … Representation
Database Query Table
GIS Query


Designed by ESRL in 1969
Features:
◦
◦
◦
◦
Querying relational databases
Tool for statistical analysis
Specific display tools (scale change or zooming)
Support for graphical formats






Data collection
Data representation
Data storage
Data analysis and computation
Concurrency and recovery
Query facility and optimization





Graphical display and interaction with the
user (ArcEdit)
Map editing tools (ArcPlot)
Network management (optimal paths)
(Network)
Layers overlaying (Overlay)
Underlying data structure (ArcScan)



Surface is represented as TIN or GRID
Grid has a reference to a Value Attribute
Table
TIN: List of Triangles, List of Associated
Nodes, List of Neighboring Triangles



Objects in ArcInfo:
◦ Point
◦ Arc
◦ Polygon
More complex objects: graphs, sequence of
arcs. Can be described in a specific
language.
For instance, a route can be defined within
a graph or a network, having specific
attributes, a set of roads, a city block can
also be represented.

Map queries: after a Map
is created, a different
coverage (highways,
regions, elevation, etc…)

2 modules:
◦ ArcView Spatial Analyst
◦ ArcView Network Analyst

Requests to the DB are written in a querytype language.




Designed by ESRI
Desktop GIS
Oriented toward data integration
Used for analyzing maps – 2D
◦ Back End Capability
Link to relational (SQL) databases or files
Data modeling
Base Mapping
Security and Integrity
◦ Front End Capability
Interactive, visual analytical performance
Manipulation of data sets
Presentation and output

Relational Database (SQL) Querying

Libraries,Images

Embedded, Linked, SQL Dictionaries

Data Formats

Raster and Vector Support

Format Handlers (loadable drivers)

Fixing Errors in Digitized Data

Storage Structures
◦ Spatial Objects (Points, Lines, Polygons)

Storage Structures
◦ Non spatial properties
◦ Derived Objects and Geo-Coding
Labels & Attributes
Line Geo-Coding

Spatial Analysis
◦ Buffering
◦ Generalization


Line Generalization
Overlay
Polygon
Dissolve
Polygon Dissolve

Data Themes
1.
2.
3.
4.
5.
6.
7.
8.
Definition
Data Source
Geographic Search Criterion
Query Condition
Graphical Style
Labels
Column List
Data

Page Layout and Printing






Licensed by GE Power Systems
Smaller GIS – business solution
OO development language (Magik)
Raster/vector graphics
Has intranet/internet network capability
High-level query language



Oracle
Postgre SQL
SDE

Index-aware operations:
◦ Contains, covers, inside, overlaps, touch, disjoint

Other:
◦ Union, difference, intersection, area, length

Spatial Attribute Types:

Spatial Operations, Spatial Indexing
◦ point, line string or polygon


Open source
Operations:
◦ Triangulate, scale, polymorphic (works on different
types of geometric objects)


No topological operations (adjacency is not
implemented)
No overlay operation
GIS
CAD
3D visualization
Custom
application
SDE server
Data files
ArcInfo files
DBMS (spatial,
CAD, graphics)




There are a number
of Software that is
being used for
statistical analysis of
spatial data.
S-Plus Extension for
ArcView GIS 3.2
SpaceStat (TerraSeer)
Grass GIS

Typical GIS software does not perform the
following functions satisfactory:
Robustness
Reliability
Completeness
Capability and features
Specific Spatial techniques for Autocorrelation
and analysis
◦ Spatial hypothesis testing
◦
◦
◦
◦
◦
Basic Statistics

Summary statistics

Crosstabulations

Hypothesis tests•

Probabilities, quantiles, and densities

Random number generation

Bootstrap and jackknife estimation

Power and Sample Size
Regression

Basic linear regression

Polynomial regression

Robust regression

Constrained regression

Logistic regression

Generalized linear models

Robust MM Regression

Linear regression with correlated
errors
Nonlinear regression and minimization

Nonlinear regression with correlated errors

Minimum-sum optimization for maximum
likelihood and generic optimization

Constrained nonlinear regression

Nonlinear mixed effects
Among Other functions

Mixed Effect Models

Classifications

Nonparametric regressions

Smoothing and interpolation

Multivariate analysis

Cluster Analysis

Quality Control
Graphical feature of the software

Time series charts: high-low-openclose and candlestick

Combined vertical/horizontal error bar
charts

Multiple x-y pair plots

Nonlinear curve fitting plots

2-D plots: area plots, barplots,
boxplots, density plots, dot charts,
histograms, pie charts, and more

Scatterplot extensions: scatterplot
matrices, linear fits, smooth fits, vary
symbol color and size, text as points

3-D plots: point clouds, surface plots,
contour plots, color image plots, 3-D
barplots

Time series plots

Quality control charts

Statistical model summaries and
diagnostics
Advanced Data Visualization

Easy control of axis scales

Pop-up descriptions for data values

2-D and 3-D graph palettes for easy plotting

Multiple simultaneous 3-D rotation views

2-D projections in 3-D space

Multiple graphs per page with auto-formatting

Interactive 3-D view angle specification

Interactive observation identification

Multiple, user-defined color maps
With the S+SpatialStats module:






Spatial autocorrelation a to assess global
association for the values of a variable as a
consequence of their location
Build neighbor weight matrices based on
adjacency as well as distance between sampling
units
Local spatial association can highlight clusters of
data spatially correlated
Supports Moran's I index and Geary's c measures
of correlation
Spatial regression to depict relationships between
variables for each spatial unit, given a neighbor
weight matrix
Model variables may be selected from ArcView
themes or S-PLUS data sets
SpaceStat:
http://www.terraseer.com/Spacestat.ht
ml

Despite solid indications that spatial
effects matter, much empirical work that
uses spatial data still fails to take its
distinctive characteristics into account.
Until SpaceStat, there was no
comprehensive software package that
covered a reasonable range of
techniques in spatial statistics and
spatial econometrics.

SpaceStat provides tools for the creation
of spatial weights matrices, exploratory
spatial data analysis and spatial
econometric analyses

SpaceStat was first released in 1991, and
since then was updated 4 times. In 2002,
SpaceStat joined forces with TerraSeer.
ClusterSeer 2 evaluates disease clusters and
non-disease events such as crime or sales
data. You can determine whether a cluster is
significant, where it is located, and when it
arose, providing insight into the origin,
causes, and correlates of the event.
BoundarySeer is the premier product for the
detection, description and analysis of
geographic boundaries. It detects patterns in
your data and then tests them statistically.
Aside from edge detection, most GIS do not
provide any boundary analysis techniques.
GrassGIS:
http://grass.itc.it/statsgrass/
 This is an open source,
language-based software,
which allows in-house
development of spatial tools
that can be used to manipulate
data that must be represented
by unconventional statistical
systems.
 All of its functionality is
essentially a subset of S-plus.

Rich variety of
software appeared
in the last 10 years
to deal with spatial
data, statistical
analysis, and
database queries
on complex
geometric objects.

Future directions
include webbased, serverbased distributed
database with
hierarchical data
representation and
advanced
visualization
capabilities