GIS 220 - Geographic data analysis 9/14/2021
Lecture 8
GIS220
Spatial Analysis of Points
Prof Gregory Breetzke
greg.breetzke@up.ac.za
Room 1-19, Geography
GIS 220 - Spatial Analysis of Points 1
Recap
• You know how to describe spatial data
– Descriptive stats
– ESDA
• Spatial data is related in space
– Autocorrelation (Moran’s I, etc.)
So how do we analyse points (spatial points)?
We will learn to summarize location, dispersion, and
degree of clustering of points.
GIS 220 - Spatial Analysis of Points 2
Spatial Analysis of Points 1
GIS 220 - Geographic data analysis 9/14/2021
Lecture outline
1. Requirements
2. Types of distributions
3. Spatial effect
4. Centrographic statistics
1. Mean center
2. Central feature
3. Standard distance
4. Standard deviational ellipse
5. Distance-based point pattern measures
GIS 220 - Spatial Analysis of Points 3
Planar
mapping
Events = Objectivity
proper Requirements (study area)
for point
pattern
analyses
1:1 (objects Inclusion
& events) (entities)
GIS 220 - Spatial Analysis of Points 4
Spatial Analysis of Points 2
GIS 220 - Geographic data analysis 9/14/2021
Types of distributions
– Random: any point is equally likely to occur at any location, and the position of any
point is not affected by the position of any other point.
– Uniform: every point is as far from all of its neighbors as possible: ‘unlikely to be close’.
– Clustered: many points are concentrated close together, and there are large areas that
– contain very few, if any, points: ‘unlikely to be distant’.
RANDOM CLUSTERED
UNIFORM/
DISPERSED
GIS 220 - Spatial Analysis of Points 5
Is the point pattern random?
• Is it random, uniformed, or clustered? How do we know?
Complete Spatial Randomness (CSR)
variation of intensity and interaction in space
GIS 220 - Spatial Analysis of Points 6
Spatial Analysis of Points 3
GIS 220 - Geographic data analysis 9/14/2021
Complete Spatial Randomness
The CSR process postulates two conditions:
1. The condition of equal probability.
– Any point has an equal probability of being in any
position or, equivalently, each area has an equal
chance of receiving a point.
2. The condition of independence.
– The positioning of any point is independent of the
positioning of any other point.
GIS 220 - Spatial Analysis of Points 7
Spatial effects
First Order Effect Second Order Effect
• Equal probability: • Independence:
– The influence of external or – Influence of one location on
environmental factors on nearby locations: e.g., non-
processes/points: e.g., contagious versus contagious
abundance of plants within a diseases
sub-region could depend on
soil type
GIS 220 - Spatial Analysis of Points 8
Spatial Analysis of Points 4
GIS 220 - Geographic data analysis 9/14/2021
Second-order
effect:
interaction
between
First-order
events
effect:
variation
in its
intensity
over space
GIS 220 - Spatial Analysis of Points 9
Describing a point pattern
First-order or second order effect?
First-order effect: variation in its intensity over space
Second-order effect: interaction between events
Spatial Analysis of Points 5
GIS 220 - Geographic data analysis 9/14/2021
What causes spatial dependence?
• Underlying socio-economic process has led to clustered
distribution of variable values.
– Grouping processes
• Grouping of similar people in localised areas.
– Spatial interaction processes
• People near each other more likely to interact, share.
– Diffusion processes
• Neighbours learn from each other.
– Dispersal processes
• People move, but generally over short distances, take their knowledge
with them.
– Spatial hierarchies
• Economic influences that bind people together.
GIS 220 - Spatial Analysis of Points 11
Describing a point pattern
• Use spatial centrographic statistics.
• Most basic kind of descriptor for spatial
distributions:
1. Mean centre
2. Central feature
3. Standard deviational distance
4. Standard deviational ellipse
GIS 220 - Spatial Analysis of Points 12
Spatial Analysis of Points 6
GIS 220 - Geographic data analysis 9/14/2021
Mean centre
n xi n yi
s = ( x , y ) = i =1 , i =1
n n
s1 (x1, y1)
GIS 220 - Spatial Analysis of Points 13
Central feature
• Identifies the most centrally located feature in
a point, line, or polygon feature class.
GIS 220 - Spatial Analysis of Points 14
Spatial Analysis of Points 7
GIS 220 - Geographic data analysis 9/14/2021
Standard Distance
• A measure of how dispersed the events are
around their mean centre
( x − ) 2 + ( yi − y ) 2
n
d= i =1 i x
GIS 220 - Spatial Analysis of Points 15
Standard Distance
• Assesses measure of dispersion and alignment.
• One SD will cover approximately 68% of all input features.
GIS 220 - Spatial Analysis of Points 16
Spatial Analysis of Points 8
GIS 220 - Geographic data analysis 9/14/2021
Standard deviational ellipse
GIS 220 - Spatial Analysis of Points 17
Horn, C. A. & Breetzke, G. D. (2009). Informing a crime strategy for the FIFA 2010 World Cup: a case study
for the Loftus Versfeld stadium in the city of Tshwane, South Africa, Urban Forum, 20(1), 19-32.
Spatial Analysis of Points 9
GIS 220 - Geographic data analysis 9/14/2021
Quantifying a point pattern
• Distance-based point pattern measures:
– Nearest neighbour distance
– Average nearest neighbour
– G function
– Nearest neighbour index (NNI)
GIS 220 - Spatial Analysis of Points 19
Nearest Neighbour Distance
• Distance from an event to its nearest neighbour
Euclidean distance
d ( si , s j ) = ( xi − x j ) 2 + ( yi − y j ) 2
Mean nearest neighbour distance
n
d min ( si )
d min = i =1
n
GIS 220 - Spatial Analysis of Points 20
Spatial Analysis of Points 10
GIS 220 - Geographic data analysis 9/14/2021
Nearest Neighbour Distance
• If clustered, d min has a higher or lower value?
𝑑ҧ𝑚𝑖𝑛 = 20,97 GIS 220 - Spatial Analysis of Points 21
G Function
Examine the cumulative frequency distribution
no.[d min ( si ) d ] of the nearest-neighbour distance.
G (d ) =
n
GIS 220 - Spatial Analysis of Points 22
Spatial Analysis of Points 11
GIS 220 - Geographic data analysis 9/14/2021
G Function
Shortest
• Shortest nearest neighbour distance is 9.
• 9 is the shortest nearest neighbour for 2
events.
• 2/12 = 0,167.
Next shortest
• Next shortest nearest neighbour is 15,64.
• There are 3 events at this distance or
closer.
• 3/12 = 0,25.
Next shortest
• Next shortest nearest neighbor is 21,14.
• There are 5 events at this distance or
closer.
• 5/12 = 0,42.
GIS 220 - Spatial Analysis of Points 23
Nearest Neighbour Index (NNI)
• Measures distance between each feature
centroid and its nearest neighbour centroid
location.
NNI=2D (N/A)
D: avg. distance between each point
𝑑ҧ𝑚𝑖𝑛 = 20,97
& its nearest neighbour
N: number of studied points
A: size of the studied area
GIS 220 - Spatial Analysis of Points 25
Spatial Analysis of Points 12
GIS 220 - Geographic data analysis 9/14/2021
Regular, random or clustered?
GIS 220 - Spatial Analysis of Points 26
NNI – Example 1 (flat land)
Settlement Nearest Neighbour Distance (km)
Cierny Majer Kosuty 1.6
Kosuty Cierny majer 1.6
Hed Cierny Brod 0.9
Cierny Brod Hed 0.9
Mostova Cierny Brod 2
Cierna Voda Cierny Brod 2.4
Stary Haj Vozokany 3
Vozokany Stary Haj 3
Degessky majer Dolna luka 3.4
Dolna luka Matuskovo 2.5
Matuskovo Budic 2.2
Budic Matuskovo 2.2
N = 12 d = 25.7
GIS 220 - Spatial Analysis of Points 27
Spatial Analysis of Points 13
GIS 220 - Geographic data analysis 9/14/2021
NNI – Example 1
• Total sum of distances = 25.7 (d)
• Area of study site = 100 km2 (hypothetical) (A)
• Number of villages = 12 (N)
• NNI = 2D (N/A)
– 2D = 2 (d/N)
= 2 (25.7/12) NNI ~ 0: clustered
= 2 * 2.14 NNI ~ 1: random
NNI ~ 2,25: regular/uniform
= 4.28
• NNI = 4.28(12/100)
= 1.48
→ regular
• A lot of free land on flat plains.
GIS 220 - Spatial Analysis of Points 28
NNI – Example 2 (hilly land)
Settlement Nearest Neighbour Distance
(km)
Mlyniste Slace 0,2
…
Koborno Kraje 0,5
Kraje Klizske Lucky 0,5
Klizske Lucky Kraje 1,5
Drahozicka huta Klizske Lucky 1,1
N = 23 = 16,1 (d)
GIS 220 - Spatial Analysis of Points 29
Spatial Analysis of Points 14
GIS 220 - Geographic data analysis 9/14/2021
NNI – Example 2
• Total sum of distances = 16,1 (d)
• Area of study site = 100 km2 (hypothetical) (A)
• Number of villages = 23 (N)
• NNI = 2D (N/A)
– 2D = 2 (d/N)
= 2 (16,1/23) NNI ~ 0: clustered
= 2 * 0,7 NNI ~ 1: random
= 1,4 NNI ~ 2,25: regular/uniform
• NNI = 1,4 (23/100)
• NNI = 0,67
→ clustered
• Valleys
GIS 220 - Spatial Analysis of Points 30
Other Applications
Wildlife distributions in Kenya.
Mineral deposits in
Kambalda, WA.
Tuberculosis Cases v Crowding.
Vegetation
fires in
India.
Vehicular crashes in Honolulu.
GIS 220 - Spatial Analysis of Points 31
Spatial Analysis of Points 15
GIS 220 - Geographic data analysis 9/14/2021
In Summary
• Centrographic statistics (mean center, standard deviation distance,
and standard deviation ellipse) measure centrality and dispersion.
– Useful for describing spatial patterns in the data.
– Useful for comparing two distributions.
• NNI is an indicator for clustering.
• Spatial autocorrelation (global: Moran’s I, local: LISA) looks at how
similar are those values that are closer to each other.
– LISA will also tell you where clustering occurs and if there are low
clusters, high clusters, as well as areas of dispersion (high-low/low-
high clusters).
GIS 220 - Spatial Analysis of Points 32
References
Briggs, R. 2010. Spatial Autocorrelation:
Introduction to Concepts. Henan University.
Getis, A. and Ord, J.K. 1992. The analysis of
spatial association by use of distance statistics.
Geographical Analysis, 24(3) 189-206.
GIS 220 - Spatial Analysis of Points 33
Spatial Analysis of Points 16
