INTRODUCTION TO SPATIAL
INTRODUCTION TO SPATIAL
ANALYSIS
Four fundamental
functions of GIS
fall
under the
manipulation and analysis
component (Martin, 1991):
1. Reclassification operations 2. Overlay operations
3. Distance and connectivity measurements 4. Neighbourhood characterisation
1.
Reclassification operations
• transform the attribute information associated with a single map coverage. E.g.
• allow the “cause-and-effect” of certain spatial factors be evaluated. E.g.
* population densities classified into classes such
as 'sparsely populated' or 'overcrowded' etc.) * soil types and farmland values
MAP
URBAN
RESIDENTIAL INDUSTRIALURBAN
RURAL FOREST
RURAL AGRICULTURE
URBAN
RURAL
ORIGINAL CLASSIFICATION GENERALIZED CLASSIFICATION
classification provides new patterns/relationships
Classification (cont.)
Classification:
• Land parcels for housing are classed into
• single storey terrace,
• double storey terrace and • bungalow.
Generalisation:
2. Overlay operations
Nb rMean of residuals - 500m.shp -0.74 - -0.57 (great est overestimation) -0.57 - -0.35
-0.35 - 0 0 0 - 0.35 0.35 - 0.57
0. 57 - 0.78 (greatest und erestim ation) No Data
Glasgow City Council - UKBORDERS.shp Motorway12km.shp
4 0 4 Miles
N
E W
S
Ordnance Survey Crown Copyright. All rights reserved
3. Distance and connectivity
measurements
• include both simple measures of inter-point
distance and more complex operations such as the construction of zones of increasing transport cost away from specified locations.
• Distance measurement can be used to calculate straight line and network distance.
MEASUREMENT MEASUREMENT DISTANCE DISTANCE PARAMETER PARAMETER AREA/SIZE AREA/SIZE A B C D
X 5 KM Y
A- B = 20 = 40% B- C = 20 = 40% C - D= 10 = 20%
2
DISTANCE (STRAIGHT LINE) MEASUREMENT
A- B: Alor Setar - Kuala Lumpur 360 KM
B- C: Kuala Lumpur - Kuantan 270 KM
4. Neighbourhood characterisation
• involves ascribing values to location according to characteristics of the surrounding region.
• Such operations may involve both summary and mean measures of a variable.
Mean selling price (£ per sq km) .shp 20500 - 45939
45939 - 71377 71377 - 96816 96816 - 122255 122255 - 147694 147694 - 173132 173132 - 198571 198571 - 224010 224010 - 249449 No Data
Glasgow City Council - UKBORDERS.shp Motorway12km.shp
4 0 4 Miles
N E W
S
Ordnance Survey Crown Copyright. All rights reserved
Neighbourhood analysis of mean selling prices
• Anselin (1998) proposes that GIS functions
can be classified as follows
– Selection
– Manipulation – Exploration – Confirmation
GIS
functions
• Selection: involves boolean queries and spatial
sampling. This seems similar to the overlay operations function.
• Manipulation: may be based on attribute data, map
data, or integration of both, simultaneously. This
means analysing data in an integrated manner where various data as available in the database can be
combined in an analysis.
• Exploration: for investigation of spatial structure and
involves description and visualisation. This is relevant to spatial autocorrelation analysis of hedonic models using geo-statistical method
• Confirmation: for modelling spatial association and/or
DATA RECALL
DATA RECALL
• can be invoked on spatial and attribute components
• involves selective search • no new objects created • examples:
CLASSIFICATION AND GENERALISATION
CLASSIFICATION AND GENERALISATION
• classification - identify a set of characteristics to group together objects.
• in a vector system, classification involves addition of objects characteristics.
• in a raster system, classification involves converting or coding cell values.
• classification examples: Land parcels for housing are classed into single storey terras, double storey terras and bungalow.
Housing Age
Legend
Before 1900 1901-1930
1931-1950
1951 to 1999
URBAN
RESIDENTIAL INDUSTRIALURBAN
RURAL FOREST
RURAL AGRICULTURE
URBAN
RURAL
...
...
• vector data
– converting attribute values for polygon, line and point
• raster data
MEASUREMENT
• measurement functions includes distance, parameter and area
• example: land parcels larger than 5 hectares
MEASUREMENT MEASUREMENT DISTANCE DISTANCE PARAMETER PARAMETER AREA/SIZE AREA/SIZE A B C D
X 5 KM Y
A- B = 20 = 40% B- C = 20 = 40% C - D= 10 = 20%
2
DISTANCE MEASUREMENT
A- B: Alor Setar - Kuala Lumpur 360 KM
B- C: Kuala Lumpur - Kuantan 270 KM
... MEASUREMENT
... MEASUREMENT
• vector data
– area and parameter is obtained from coordinates of the polygon nodes
– distance is derived from coordinates of starting/ending nodes
SEARCHING
SEARCHING
• determine values against target object according to a neighbourhood characteristic
• three parameters need to be identified
– targets
– neighbourhood around the targets
– applied neighbourhood function for resultant values
• example: total of households within 1 km of proposed shopping mall
– target-shopping mall
... SPATIAL SEARCH
... SPATIAL SEARCH
• operated as additional points in polygon, line in polygon and polygon in polygon
• vector data
– point, line or polygon analysed with
neighbourhood polygon using coordinate nodes
– involves complex calculation with overlapping and out-of-boundary neighbourhood
• raster data
NEIGHBOURHOOD
NEIGHBOURHOOD
• represents ‘distance’ between map features • ‘distance’ unit can be in measurement units
or other units like travelling time, noise level, visibility distance etc.
• requires 4 parameters
... NEIGHBOURHOOD
... NEIGHBOURHOOD
• used to generate buffer zones
• example: a 2km zone along a proposed
transmission line alignment; zones exceeds 50dB around the airport
• neighbourhood is most often complex and involves data from various layers. For
R Buffering a Point
eg. All area within one mile of a city
Buffering a Line
eg. All areas within 100 meters of a road
Buffering an Area
OVERLAY
OVERLAY
• Involves two or more data layers • Produces new layers
• Two types of overlay operation – arithmetic overlay
– logical overlay
MAP ALGEBRA (MULTIPLICATION)
MAP ALGEBRA (MULTIPLICATION)
OVERLAY BY MULTIPLICATION
DISTRICT
1 2
3 4 X
CROP AREA
1
B B
= 1 2
3 4
OVERLAY BY MAXIMUM VALUE
3 3 4 0 1 0 2 4 6
4 2 2
5 5 5
4 1 1
4 3 4
5 5 5
4 4 6
RAINFALL : RAINFALL: RAINFALL:
1980 1981 1980 - 1981
+ =
... OVERLAY
... OVERLAY
• vector data are sometimes more
efficient than raster data if data are not dense.
– vector data - operation based on the
selected data only
– raster data - operation on all cells -
Soil Type
+
Crops Production (ton/ha)
Overlay Result
GIS Technology: Relationship between Land use and Crop Productivity
Ja lan R
aja A lang
Jalan Abdul Aziz
Sewerage pond Sewerage pond
Ja lan R
aja A lang
Jalan Abdul Aziz
Ja la n D at uk M al ik Jalan Hamzah
Jalan Ra
ja Uda
Ja lan R
aja A lang
Jalan Abdul Aziz Jalan Hamzah Jalan Raja Uda