Descriptive Spatial Analysis

1
Descriptive Spatial Analysis

• Definition of Crime Mapping
• Single Symbol Mapping
• Buffers
• Chart Mapping
• Graduated Mapping
• Hotspot Analysis
• Practical Examples

2
Definition of Crime Mapping

• Crime analysis is
• A geographic information system (GIS) is a set of
  computer-based tools that allow a person to
  modify, visualize, query, and analyze geographic
  and tabular data.
• Consequently, computerized crime mapping is the
  process of using a geographic information system
  in combination with crime analysis techniques to
  focus on the spatial context of criminal and
  other police activity.

Source Boba, R. (Forthcoming). Crime mapping. In
Encyclopedia of criminology. Chicago Fitzroy
Dearborn Publishers.
3
Single Symbol Mapping

• Uses individual symbols to represent point, line,
  and polygon features.
• Allows for a detailed analysis of small amounts
  of data.

4
Example Too Much Data
5
Example Tabular Data
6
Example Geographic Data
7
Example Geographic Data
8
Buffers

• A buffer is a zone of a specified distance around
  a feature.
• Points, lines, and polygons can be buffered.
• Buffers are useful for proximity analysis and can
  be designated at one or many intervals (e.g., 500
  feet, 1,000 feet, 1 mile).

9
Buffers Point Example
10
Buffers Line Example
11
Buffers Polygon Example
12
Chart Mapping

• A chart map allows for the display of the values
  of many data attributes at once with either a pie
  or a bar chart.
• The mapping program takes the values for numerous
  variables and displays them in a pie or a bar
  chart on the designated location on the map.

13
Chart Mapping Pie Chart Example
14
Chart Mapping Bar Chart Example
15
Graduated Size Mapping

• Data are summarized so that symbols (point or
  line features) are altered in size to reflect the
  frequencies in the data.
• Reflect more incidents at a given location with a
  larger symbol or a thicker line.

16
Example Too Much Data
17
Graduated Size Point Mapping Example
18
Graduated Size Line Mapping Example
19
Graduated Color Mapping

• Point, line, or polygon features are shaded
  according to a statistical formula, custom
  setting, or unique value. Also called choropleth
  mapping.
• Most Commonly Used Unique Value, Natural Breaks
  (default), Custom.
• Others Quantile, Equal Area, Equal Interval,
  Standard Deviation.

20
Points Shaded by Unique Value Geographic Data
21
Points Shaded by Unique Value Geographic Data
22
Points Shaded by Unique Value Tabular Data
23
Points Shaded by Unique Value Tabular Data
24
Natural Breaks

• The default classification method in most GIS
  programs.
• Identifies natural break points between classes
  using a statistical formula.

Graduated Polygon Example
25
Custom

• Ranges can be determined by the user and are not
  based on the data.
• Important for comparing the same type of data
  over time.

Graduated Polygon Example
26
Quantile

• Each class contains the same number of features
  (data points).

Graduated Polygon Example
27
Equal Interval

• Divides the range of attribute values into equal
  sized sub-ranges.
• Features are then classified based on the
  sub-ranges.

Graduated Polygon Example
28
Standard Deviation

• The GIS determines the mean value and then places
  class breaks above and below the mean based on
  the standard deviation.

Graduated Polygon Example
29
Use of Classifications

• Classifications are the descriptive statistics of
  spatial analysis. Thus, they should be
  controlled by the analyst and carefully applied.
• A danger is that the GIS has defaults (natural
  breaks into five categories) and analysts do not
  change them.
• Guidelines
• Use most, if not all, of the classifications in
  the beginning of the analysis to determine the
  nature of the data and its distribution.
• Experiment with number of categories and
  classifications to see how the maps change.
• Determine the purpose of the analysis and choose
  the best classification.

30
Exercise

• Scenario
• You are a member of a problem-solving team tasked
  with addressing an ongoing robbery problem in the
  city. You have been asked to bring an analysis
  of robbery to the first meeting. What type of
  map would you bring?
• How much data?
• Which unit of analysis?
• Which classification?

31
Graduated Points
32
Graduated Color Polygons Natural Breaks
33
Graduated Color Polygons Standard Deviation
34
Exercise

• Scenario
• As part of an impact evaluation for a problem
  analysis project to reduce commercial burglary,
  you are asked to prepare a map that compares
  before and after (same amount of time) the
  response by block group.
• How would you present this in two maps?
• In one map?

35
First of two maps
36
Second of two maps
37
In one map Difference between Pre and Post
38
Exercise

• Scenario
• The chief asks you to examine aggravated assault
  and simple assault in the city to see if there
  are differences in the relative frequencies by
  block group (or other polygon). That is, are
  there some areas that are higher in aggravated
  assault than others and are those the same that
  are higher in simple assault?

39
Using Standard Deviation Aggravated Assault
40
Using Standard Deviation Simple Assault
41
Using Quantile Aggravated Assault
42
Using Quantile Simple Assault
43
Hotspot Analysis

• In this context, the term hotspots refers to
  concentrations of events confined to a particular
  geographic area that occur over a specific time
  period. Hotspots are also referred to as
  clusters or concentrations.
• Methods for determining hotspots
• Graduated color maps
• Map grids
• Ellipses
• Kernel density interpolation

44
Hotspot Analysis

• Graduated Color Maps
• Point, line, or polygon features are shaded
  according to a statistical formula, custom
  setting, or unique value.
• In this example, census groups are shaded by the
  number of incidents.
• Note incidents are placed on the map at their
  address location for reference.

45
Hotspot Analysis

• Map grids
• Each grid cell is shaded according to the number
  of incidents.
• Unlike the preceding graduated color map, this
  method allows for smaller search areas.
• However, the grids are arbitrary and may not
  depict realistic separation of land areas.

46
Hotspot Analysis

• Ellipses
• Ellipses are drawn around the most dense
  concentrations of activity.
• Software such as S.T.A.C. (Spatial and Temporal
  Analysis of Crime), developed by the Illinois
  Criminal Justice Information Authority (ICJIA),
  uses a statistical method to find clusters.

2nd order cluster
1st order clusters
47
Hotspot Analysis

• Kernel Density Method
• A grid is applied to the map, and a score is
  derived based on the number of incidents within
  each grid cell as well as the distance to other
  incidents.
• Cell size and search radius can be dictated by
  the user.

48
Hotspot Analysis

• Factors to consider
• Definition of a hotspot
• Choice of variables
• Number of hotspots
• Scale
• Grid size and search area
• Visual display
• Comparisons
• There are many different methods of hotspot
  analysis, and each technique will reveal
  different groupings and patterns within the
  groups.

49
Practical Examples of Descriptive Mapping
50

• To assist in resource allocation of ATF agents
  analysis of gun tracing incidents per county for
  numerous states.

51

• To assist in resource allocation of ATF agents
  analysis of number of agents per county for
  numerous states.

52

• To assist in resource allocation of ATF agents
  analysis of gun tracing incidents and number of
  agents per county for numerous states.

53
Problem Analysis Project Discussion
54
Local Level Risk Assessmentfor Homeland Security

• Various geographic data are used in combination
  to assign a score to an area. The score is a
  combination of values (weighted) that can be
  based on either the presence/absence of features.
  
• The result is a thematic shading of polygons with
  the darkest (highest score) implying a higher
  risk. (Note that there is no probability
  assigned, only a score.)
• This method can be used for other types of crime
  (e.g., risk of auto theft, robbery, etc.)

55
Features to Consider

• Nuclear power plants
• Ammonium nitrate repositories
• Airports
• Amtrak
• Mass transit lines
• Amusement parks
• Malls
• Hydro Plants
• Landmarks

• Research laboratories
• Dams
• Petroleum refineries
• Ports
• Government buildings
• Interstates
• Rivers
• Population levels
• Major utility lines
• Etc.

56
Example
57
Zones containing some part of a government
building or property (note polygons).
58
Zones through which rivers flow.
59
Zones that border railroads.
60
Zones that contain schools.
61
All Zones within ½ mile of major research
facilities (weighted).
62
Total Risk Assessment
63
Alternative method Using arbitrary grids (same
sized area).
64
Caution

• This method is not tested, and many decisions are
  subjective (e.g., what data to include, values
  given to the variables).
• Also
• What should the unit of analysis be? Beat?
  Grid?
• If an arbitrary grid, what should the grid area
  be? What should the grid cell size be?
• Which of the many types of data available should
  be included and when? (Different jurisdictions
  will include different types of data.)
• How should the variables be scored in relation to
  one another? For example, should nuclear
  facilities be weighted more than malls?

65
Problem Analysis Project Discussion