Analyzing Geographic

Distributions and Point Patterns

 Analyzing Geographic Distributions:
Centro-graphy
• Centro-graphic statistics are tools used to
analyze geographic distributions by measuring
the center, dispersion and directional trend of a
spatial arrangement.
 Analyzing Geographic Distributions:
Centrography
• Mean Center
• Median Center
• Central Feature
• Standard Distance
• Standard Deviational Ellipse
• Locational Outliers and Spatial Outliers
 Mean Center

• Mean center is the geographic center for a set

of spatial features. It is a measure of central
tendency and is calculated as the average of
the xi and yi values of the centroids of the
spatial features.
• The mean center can be calculated
considering weights
Median Center
• Median center is a point that minimizes the
travel cost from the point itself to all other
points in the dataset. It is a measure of central
tendency, calculated as shown in
Central Feature

• Central feature is the object with the

minimum total distance to all other features. It
is a measure of central tendency.
Standard Distance
• Standard distance is a measure of dispersion
(spread) that expresses the compactness of a
set of spatial objects.
• It is represented by a circle the radius of which
equals the standard distance, centered on the
mean center of the distribution.
 Standard Deviational Ellipse
• Standard deviational ellipse is a measure of
dispersion (spread) that calculates standard
distance separately in the x and y directions.
Locational Outliers and Spatial Outliers
• A locational outlier is a spatial object that lies
far away from its neighbors.
• One simple method is “an object whose
distance to its nearest neighbor exceeds 2.5
deviations from the mean nearest neighbor
average (computed for the entire dataset) is
considered the locational outlier, as in
 Geo-marketing analysis:
• Locational outliers might not be desirable. For example,
if a bank wants to locate a new branch, the existence of
locational outliers might place the location point in an
area that will not be convenient for most of its clientele.
• The temporary exclusion of the locational outlier will
allow for a better branch location. The outlier will then
be included in the dataset for other analyses. In other
words, when we exclude an outlier, we usually exclude
it only for certain spatial statistical procedures, but we
include it again because it might be helpful for another
type of analysis.
 Crime analysis:
• Identifying if locational outliers in crime incidents exist may
reveal abnormal, unexpected behavior that might need
additional surveillance. For example, if a credit card that is
usually used in specific Mean center Mean center
• (A) Mean center without any locational outlier.
• (B) Mean center when the outlier is included in the calculation.
The mean center is shifted toward the direction of the
locational outlier, distorting the real center of the distribution.
160 Analyzing Geographic Distributions and Point Patterns
locations of a city (e.g., shops, restaurants, a house) is also
used the same day at a location miles away, that would be a
strong indication of fraud.
• (A) Mean center without any locational outlier.
• (B) Mean center when the outlier is included in the calculation. The mean
center is shifted toward the direction of the locational outlier, distorting
the real center of the distribution
 Analyzing Spatial Patterns: Point Pattern
Analysis
• A spatial outlier is a spatial entity whose non-
spatial attributes are considerably different
from the non-spatial attributes of its
neighbors
Spatial outlier detection can be used for the
following:
• Crime analysis: One could identify if a postcode
has a high crime rate while all the neighboring
ones have low rates. This might indicate a ghetto.
• Geo-marketing analysis: Identifying a specific
postcode where people buy a certain product
considerably less often than in adjacent ones
might indicate where access to the product is not
easy or where a better marketing campaign is
required.
 Analyzing Spatial Patterns: Point Pattern
Analysis
• Spatial point pattern S is a set of locations S =
{s1, s2, s3,...sn}
• Event is the occurrence of a phenomenon, a
state, or an observation at a particular
location.
• Spatiotemporal point pattern is a spatial
pattern of events that evolves over time. In
spatiotemporal point patterns, multiple sets of
events occur diachronically.
 Spatial Process
Spatial process is a description of how a spatial
pattern can be generated.
• There are three main types of spatial process
1. Complete spatial randomness process
2. Competitive process
3. Aggregating process
 Spatial Arrangement/Pattern
There are three main types of spatial
arrangement/pattern associated with the above
spatial processes.
1. Random spatial pattern
2. Clustered
3. Dispersed
 First- and Second-Order Effects
• First-order spatial variation effect occurs when the
values or locations of spatial objects vary from
place to place due to a local effect of space. For
example, stroke event locations may vary from
place to place inside a city.
• Second-order spatial variation effect occurs when
there is interaction among nearby locations. Here,
the location, or the value of an observation, is
highly influenced by a nearby location or the value
of a neighboring observation.
 Stationary Process
• First-order stationary process
• Second-order stationary process
• Intensity of a spatial process
• Stationary spatial process.
• Anisotropic process
• Specific directions
• Isotropic process
• Spatial heterogeneity
 Point Pattern Analysis Methods
• There are two main (interrelated) methods of
analyzing point patterns, namely the distance-
based methods and the density-based
methods.
• Distance Based
• Density Based
 Nearest Neighbor Analysis
• Nearest neighbor analysis (also called average
nearest neighbor) is a statistical test used to
assess the spatial process from which a point
pattern has been generated. It is calculated
based on the formula
 Ripley’s K Function and the L Function
Transformation
• Ripley’s K function is a spatial analysis method
of analyzing point patterns based on a
distance function. The outcome of the
function is the expected number of events
inside a radius of d.
• It is calculated as a series of incremental
distances d centered on each of the events in
turn
 Kernel Density Function
Kernel density estimation is a nonparametric
method that uses kernel functions to create
smooth maps of density values, in which the
density at each location indicates the
concentration of points within the neighboring
area.
 Lab 3 Spatial Statistics
• Overall Progress
• Scope of the Analysis: Crime Analysis
• Measuring Geographic Distributions
• Point Pattern Analysis
• Kernel Density Estimation
• Locational Outliers
 Measuring
Geographic
Distributions
Point Pattern Anaysis
 Ripley’s
Function
Kernel Density