Geog 463: GIS Workshop

1
Geog 463 GIS Workshop

• May 17, 2006
• Exploratory Spatial Data Analysis

2
Outlines

• I. Fundamentals of ESDA
• What is Exploratory Spatial Data Analysis (ESDA)?
• ESDA basics
• II. Techniques of ESDA with focus on area-class
  data
• ESDA for describing non-spatial properties of
  attribute
• ESDA for describing spatial properties of
  attribute
• III. Applications of ESDA
• Gallery of implemented ESDA systems

3

• I. Fundamentals of ESDA

4
1. What is ESDA?

• Exploratory Spatial Data Analysis (ESDA)
• Exploratory Data Analysis (EDA)
• EDA and statistics
• EDA and visualization
• EDA and cartographic visualization

5
Exploratory Spatial Data Analysis

• Extension of exploratory data analysis (EDA) to
  detect spatial properties of data
• EDA
• consists of a collection of descriptive and
  graphical statistical tools
• intended to discover patterns in data and suggest
  hypotheses
• by imposing as little prior structure as possible
• ESDA links numerical and graphical procedures
  with the map

6
Exploratory Data Analysis

• Aimed at (1) pattern detection (2) hypothesis
  formulation (3) model assessment
• Use of graphical and visual methods (e.g. Box
  plot) Use of numerical techniques that are
  statistically robust (e.g. P-value)
• Emphasis on descriptive methods rather than
  formal hypothesis testing
• Exploratory in that it cannot explain the
  patterns it reveals

7
EDA and Statistics

• Evolutions of statistics return of original
  goals of statistics in data-rich and high
  computing environment stay close to the original
  data

Image source Adrienkos website
8
EDA and Visualization

• By its very nature the main role of EDA is to
  open-mindedly explore, and graphics gives the
  analysts unparalleled power to do so
• The greatest value of a picture is when it forces
  us to notice what we never expected to see
• John W. Tukey

9
EDA and Cartographic Visualization

• Emphasis on the role of highly interactive maps
  in individual and small group efforts at
  hypothesis generation, data analysis, and
  decision-support
• Contrast with static paper maps

10
Early examples of ESDA
Dr. John Snow Investigation of deaths from
cholera London, September 1854
death locations
spatial cluster
infected water pump?
A good data representation is the key to solving
the problem
11
2. ESDA Basics

• Visual tools for non-spatial analyses
• Univariate
• Multivariate
• Visual tools for spatial analyses
• First-order properties
• Second-order properties
• Brushing Linking

12
Visual tool for non-spatial analyses

• Univariate
• Histogram
• Box plot
• Multivariate
• Scatter plot
• Parallel coordinates plot

13
Histogram, box plot
Dispersion graph
Dot plot
Distribution of attribute values within a range
Box plot
Histogram
Distribution of attribute values at y-axis given
categorical variables at x-axis
14
Scatter plot
Scatter plot shows how two attributes are related
Scatter plot matrix shows how a set of two
attributes are related
15
Parallel coordinates plot
Parallel coordinates plot object characteristics
profiles relationships between attributes (look
at line slopes)
16
Visual tools for spatial analyses

• First order properties
• Tools for exploring general trends
• Spatially lagged boxplot
• Kernel estimation
• Second order properties
• Tools for exploring spatial autocorrelation
• Moran plot

17
Spatially lagged boxplot

• Boxplot in which the categorical variable is
  spatial lag order (as defined by spatial weight
  matrix)
• After the user has selected an origin zone, a
  sequence of box plots (one for each lag order) is
  generated at increasing distance from the origin
  zone up to a user specified maximum

18
Wise et al 1998
19
Kernel Estimation

• This method is used to smooth a given point
  pattern such as crime locations so that we can
  easily detect hot spot.

20
(No Transcript)
21
Moran plot

• A plot of attribute value on the vertical axis
  against the average of the attribute values in
  the adjacent areas using spatial weight matrix
• A scatter of values sloping upward to the right
  is indicative of positive autocorrelation

22
(No Transcript)
23
Brushing linking

• Brushing a subset of data is selected and
  highlighted
• Linking map and graph are linked such that
  multiple views are displayed

Image source Symanziks website
24

• II. Techniques of ESDA

25
3. ESDA for describing non-spatial properties of
attribute

• Median
• Measure of the center of the distribution of
  attribute values
• ESDA queries which are the areas with attribute
  values above (below) the median?
• Quartile and inter-quartile spread
• Measure of spread of values about the median
• ESDA queries which are the areas that lie in the
  upper (lower) quartile?
• Box plots
• Graphical summary of the distribution of
  attribute values
• ESDA queries where do cases that lie in specific
  parts of the boxplot occur on the map? Where are
  the outlier cases located on the map?

26
(No Transcript)
27
4. ESDA for describing spatial properties of
attribute

• Smoothing
• Identifying trends and gradients on the map
• Spatial autocorrelation
• Detecting spatial outliers

28
Smoothing

• Smoothing may help to reveal the presence of
  general patterns that are unclear from the mosaic
  of values
• ESDA techniques spatial averaging take the
  attribute value of an area and its neighbors and
  average them repeat for each area

29
Identifying trends and gradients on the map

• Are there any general trends or gradients in the
  map distribution of values?
• ESDA techniques include
• Kernel estimation
• Taking transects through the data and plotting
  with attribute value on vertical axis and spatial
  location on horizontal axis
• Spatially lagged boxplot with lag order specified
  with respect to a particular area or zone

30
Spatial autocorrelation

• Propensity for attribute values in neighboring
  areas to be similar
• ESDA techniques include
• Moran plot

31
Detecting spatial outliers

• An individual attribute value is not necessarily
  extreme in the distributional sense but is
  extreme in terms of the attribute values in
  adjacent areas
• ESDA technique run a linear squares regression
  on the Moran plot, and select cases significantly
  deviated from the regression line

32
(No Transcript)
33

• III. Applications of ESDA

34
5. Gallery of ESDA systems

• GeoDa
• https//www.geoda.uiuc.edu/default.php
• CommonGIS
• http//www.commongis.com/

35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
Interactive map symbolization in CommonGIS
By moving the slider, we see more patterns and
gain more understanding of value distribution
Porto
Lisboa
Clusters of low values around Porto and Lisboa
Clusters of high values in central-east
One more cluster of low values
Coast-inland contrast
West-to-east increase
43
Link between information visualization techniques
and maps
Map and dot plot each district shown on the map
is also represented by a dot
Map and scatter plot the same technique
Map
Dot plot
A district pointed on the map with the mouse is
simultaneously highlighted on the map and the plot
44
Using Cumulative Curves
Some statistics about the result
In the most part of Portugal (coloured in blue)
the proportion of people having high school
education is below 4.67. However, on this large
territory only one third of the countrys
population lives.
In these areas over 7.82 people have high school
education. Here lives 33.1 of the total
countrys population.
45
Focusing multiple views
and here,
An object pointed on the map with the mouse
and here,
but not here this is an aggregated view that
does not show individual objects
is simultaneously highlighted here,
46
Focusing and Visual Comparison on Other Map Types
Outlier Maximum represented value Value to
compare with Minimum value
47
Spatial Distribution of Events
The small circles represent the earthquakes that
occurred in Western Turkey and the neighbourhood
between 01.01.1976 and 30.12.1999
Here we see only the earthquakes that occurred
during 30 days from 15.05.1977 to 13.06.1977
By applying the temporal filter, we can
investigate the spatial distribution on any time
interval
48
Progress of Spatial Patterns over Time
Map animation allows us to see how the spatial
distribution of events and their characteristics
evolve over time
25.05.1977 - 23.06.1977
04.06.1977 - 03.07.1977
15.05.1977 - 13.06.1977
14.06.1977 - 13.07.1977
04.07.1977 - 02.08.1977
24.06.1977 - 23.07.1977
Each animation frame in this example covers
30-days time interval. The step between the
frames is 10 days. Hence, there is 20 days
overlap between the adjacent frames.
49
Exploration of Behaviors
The value flow symbols show us the evolution of
attribute values (behavior) at each location.
Unfortunately, symbol overlapping creates
significant inconveniences, and zooming does not
always help
50
Data Transformations for Behavior Exploration
As with time maps, various data transformations
can be applied to value flow maps.
Here we have applied the comparison to the mean
the values for each moment are replaced by their
differences to the countrys mean at the same
moment. Yellow colour corresponds to positive
differences, and blue to negative. We have
received a rather clear spatial pattern.
51
Summary Characteristics of ESDA
High interactivity
Due to direct manipulation computer screens will
play no less revolutionary role for data
exploration than the invention of Cartesian
coordinates
W.Cleveland 1993
Enabling multiple complementary views
allow the user ... to see data from multiple
perspectives
A.MacEachren and M.-J. Kraak 1997
52
Summary Methods of ESDA

• Manipulating data
• Varying the symbolization
• Manipulating the users viewpoint
• Highlighting portions of a data set
• Multiple view
• Animation
• Linking maps with other forms of display
• Access to miscellaneous resources
• Automatic map interpretation (i.e. data mining)

From Slocum et al 2005
53
Discussion questions

• Assess the value of ESDA techniques in analyzing
  any geographical data with which you are familiar
• Discuss the strengths and weakness of current GIS
  software for undertaking ESDA

54
Value of ESDA in analyzing spatial data

• Help reveal unknown pattern that couldnt be
  revealed without multiple views or other ESDA
  mechanisms
• Moran plot for identifying spatial outlier
• Parallel coordinate plot for looking at the data
  distribution of a particular record relative to
  other records
• Help create a map that fits into users need
• Can select a subset of data related to map
  purpose (user interaction)
• Help avoid jumping to the conclusion with a
  single thematic map or solely based on visual
  impact
• By letting users explore the consequence of
  different map symbolization or map design
• By letting users determine whether the pattern is
  unusual (use of statistics)

55
Weakness and strength of GIS for undertaking ESDA

• CAN
• Identify smooth properties
• Techniques for describing non-spatial property of
  attribute (e.g. mean, standard deviation)
• Presentation graphics maps, histogram
• CANT
• Identify rough properties (e.g. outliers or
  spatial outliers)
• GIS has stronger PRESENTATION components than
  EXPLORATION components GIS was not originally
  designed to data exploration

56
References

• Anselin, 1998, Geocomputation A Primer, pp.
  77-94
• Anselin, 2005, GeoDa workbook
• Haining Wise, 1998, Providing scientific
  visualization for spatial data analysis criteria
  and assessment of SAGE, retrieved from
  http//www.ersa.org/ersaconfs/ersa98/papers/409.pd
  f
• Haining Wise, 2000, GISCC Unit 128
• Slocum et al, 2005, Thematic Cartography and
  Geographic Visualization, pp. 389-405
• Wise et al, 1998, The role of visualization in
  the exploratory spatial data analysis of
  area-based data, retrieved from
  http//www.geocomputation.org/1998/81/gc_81.htm
• Adrienkos website http//www.ais.fraunhofer.de/a
  nd/
• One of authors of CommonGIS
• Symanziks website http//www.math.usu.edu/syman
  zik/
• One of authors of xGobi