advanced spatial analysis - christina friedle

ADVANCED SPATIAL ANALYSIS

GIS Analysis | Fall 2013

There are many types of Spatial Analysis (Vector &

Raster). This lecture is focused on the following types

of Advanced Analysis:

Spatial Descriptive Statistics

Spatial Interpolation

Image Classification

Surface Analysis

Optimization

1st Law of Geography

Waldo Tober’s 1st Law of Geography

"Everything is related to everything else, but near things

are more related than distant things.”

Spatial Autocorrelation

A measure of the degree to which a set of spatial

features and their associated data values tend to be

clustered together in space (positive spatial

correlation) or dispersed (negative spatial

autocorrelation)


Spatial equivalent of the descriptive statistics

commonly used in statistical analysis, e.g. mean and

standard deviation.

Non-Spatial Statistics

Attribute Table “Statistics” tool

Non-Spatial Statistics

Cell Statistics tool

ArcToolbox > Spatial Analyst > Local > Cell Statistics

Common Spatial Descriptive Statistics

Nearest Neighbor index

Spatial Autocorrelation (Moran’s I)

Mean Center

K-means algorithm

Thiessen (Voronoi) polygons

Nearest Neighbor Index

How clustered or dispersed are features?

Compares observed average distance from features

to nearest neighboring features vs. expected

average distance (if random dispersion). Looks for

spatial patterns in features.

ArcToolbox > Spatial Statistics > Analyzing Patterns >

Average Nearest Neighbor


Output of Nearest Neighbor Index is an index value

and related statistics; can be viewed in the

Geoprocessing “Results” window, or used in analysis

models/scripts.

Spatial Autocorrelation Index (Moran’s I)

Measures the similarity among feature attributes

relative to feature locations. Looks for spatial patterns

in attributes.

ArcToolbox > Spatial Statistics > Analyzing Patterns > Spatial Autocorrelation

(Morans I)


Output of Spatial Autocorrelation is also an index

value and related statistics accessed via “Results”

window.

What is a “Centroid”?

Point representing the center of a feature, or of a

group of features (e.g. multipoints, lines, or

polygons).

Mean Center

Average of all input points or centroids (based on

x/y values). Output is a single point.

ArcToolbox > Spatial Statistics >

Measuring Geographic Distributions > Mean Center

Mean Center

K-means

An iterative (repeating) algorithm that finds

geographic groups of point features and determines

their mean centers (e.g. store locations). Output is a

set of points.

Thiessen (Voronoi) Polygons

Each thiessen polygon contains a single point. Any

location within a polygon is closer to its associated

point than any other point feature (e.g. store

coverage areas). Output is a set of polygons.

ArcToolbox > Analysis > Proximity > Create Thiessen Polygons


Estimating values of a continuous representation in

places where the values have not been measured

In ArcGIS, spatial interpolation tools generate a

new raster dataset

Spatial Interpolation Methods

Spline

Inverse-distance weighting (IDW)

Kriging

Density Estimation

Interpolation | Spline

Interpolates a surface from a set of points using a

“minimum curvature spline” technique; conceptually

like bending a sheet of rubber to pass through the

points.

Interpolation | Inverse-distance weighting (IDW)

Estimates unknown values as weighted averages of

the known measurements at nearby points, giving

the greatest weight to the nearest points (using

Tobler’s Law).

Interpolation | Kriging

Similar to the IDW method in that it applies weights

to the data based on distance;

Differs from IDW in that it also takes into account

the form and spatial structure of all the data.

Provides a measure of the

certainty or accuracy of the

prediction.

Best used when data is known

to have spatial autocorrelation

Inverse Distance Weighting (IDW) Kriging

Density Estimation

Spreads known quantities of a phenomena across

the landscape, based on quantities that are

measured at point sample locations and search

areas around each location.

(Point locations display census population figures for each town)

Population Density across USA


Extracts information from a multi-band raster image

Image Classification | Unsupervised

Finds natural groupings (clusters) of spectral classes

in a multi-band image; e.g. Island vs. Ocean

Image Classification | Supervised

Uses the spectral signatures obtained from training

samples to classify an image; e.g. vegetation classes

Image Classification toolbar:

Surface Analysis

Surface Analysis based on Elevation

Operations related to analysis of raster surfaces

Variety available in ArcToolbox:

Slope

Hillshade

Aspect

Contours

Viewshed

Watershed delineation

Slope

The incline or steepness of a surface or terrain

Can be measured in degrees (0-90) or percent

slope (rise/run)*100

Hillshade

Shadows drawn on a map to simulate the effect of

the sun’s rays over the varied terrain of the land

Creates a 3D effect that provides a visual sense of

shaded relief and a relative measure of incident

height for analysis

Aspect

The compass direction that a topographic slope

faces, usually measured in degrees from North

Viewshed

The locations visible from any given point or line

Finding well exposed places for communication lines

or hidden places for parking lots

Elevation in the area of the observation point

Green cells are visible from the observation point

Viewshed

Contour Lines

A line on a map that connects points of equal

elevation usually based on sea level (or another

vertical datum)

Watershed Delineation

Determines the contributing area above a set of

cells in a raster; a watershed is an area that drains

to a common outlet (a.k.a., basin, catchment)

Optimization

Optimization

Analytical techniques used to determine the best

(optimal) path, location, or other geographic

parameter based on a set of criteria

Common Optimization Analyses

Route (“Least Cost Path”)

Origin-Destination (OD) Cost Matrix

Service Areas

Location-Allocation

Site Suitability

Route (“Least Cost Path”) Analysis

Computes a travel route between locations along a

network, based on impedance (e.g. distance, time,

scenic beauty).

The “best” path is the one with the lowest impedance

score, or “least cost”.

Network Analyst toolbar > New Route

Origin-Destination (OD) Cost Matrix Analysis

Finds and measures distances along network least-

cost paths, from multiple origins to multiple

destinations

Network Analyst toolbar > New OD (Origin-Destination) Cost Matrix (layer)

Service Areas Analysis

Identifies areas along all paths in a network that are

within an impedance value (e.g. 5 minutes) from a

starting location.

Network Analyst toolbar > New Service Area (layer)

Location-Allocation Analysis

Locates facilities in such a way that demand for

services is allocated to each facility efficiently.

Network Analyst toolbar > New Location-Allocation (layer)

Location-Allocation Analysis

Locating an Emergency Response Center.

“Where should three ERC facilities be placed so that the

greatest number of people in the community can be

reached within 4 minutes?”

Locating a manufacturing plant.

“Where should the manufacturing plant be located to

minimize overall transportation costs?”

Site Suitability Analysis

Process of identifying the best location for a

particular feature(s) based on suitability criteria for

multiple, overlaying GIS datasets.

Criteria can be weighted.

Table 1. Bioethanol plant site selection suitability criteria and ranking values.

Suitability

Values

Low

Suitability

Medium

Suitability

High

Suitability

Assigned

Importance

Decimal

Weight

Major Roads > 1 mi 0.5 - 1 mi < 0.5 mi 3 0.1764706

Railways > 3 mi 1.0 - 3 mi < 1 mi 2 0.1176471

Towns > 3 mi 1.0 - 3 mi < 1 mi 3 0.1764706

Power Lines > 1 mi 0.5 - 1 mi < 0.5 mi 3 0.1764706

Maize Fields > 3 mi 1.0 - 3 mi < 1 mi 3 0.1764706

Rivers > 3 mi 1.0 - 3 mi < 1 mi 2 0.1176471

Airports > 5 mi 1.0 - 5 mi < 1 mi 1 0.0588235

advanced spatial analysis - christina friedle

Documents