Maps A graphicresearch tool

1
MapsA graphic/research tool
2
Visual Thinking

• Vertical Thinking
• Connections between phenomena
• Ex Neighborhood characteristics
• Horizontal Thinking
• Common factors across a particular crime
• Visual Thinking is private and develops
  hypotheses .. (educated guess)

3
Hypothetical- deductive thinking

• Steps
• Development of a hypothesis
• Development of a method for testing
• Analysis of the data.
• Evaluation of the results
• A decision to accept or reject the original
  hypothesis
• Reevaluation of the original hypothesis,

4

• Is this research model feasible day to day?
• NO .. It is an ideal
• Very few police departments have time , resources
  or skills to do original research

5
Visual Communication

• We synthesize the results to present a point
• Can we lie with maps?
• Yes
• As a society we like maps but
• We do not often have the training to interpret
  them
• How many of you ever had a geography course?

6
Map Types

• Statistical
• Point
• Choropleth
• Isoline
• Surface
• Linear

7
Each map type conveys differently

• Thematic Maps they display a specific theme
• Dot maps showing each crime
• Dot maps showing total crimes per location
• Area shading showing crimes per area .. Ward,
  neighborhood, Census area unit
• Statistical map show numbers at points or
  areas???
• Statistical maps can show graphs charts in each
  respective area .. Graduated symbols

8
Graduate Symbols
9
Proportional Symbols
10
Proportional Symbols as Pictographs
11
Dot Map Dots per Area
12
Multivariate Dot Map
13

• http//www.philly.com/inquirer/special/violence/

14
Thematic Maps Choropleth Maps

• Polygon based area shading
• Mapping total numbers using choropleth mapping
  is unacceptable
• There is a misleading impression given by unequal
  areas.
• Example Total Population by County

15
Thematic Maps Choropleth Maps

• How do we adjust for this?
• Averages means and medians
• Crime rates crimes per ????? Persons
• Density Measures Persons per mile
• Percentages percent of area in park land
• Careful interpretation crimes per population?
• Crimes in a low density population area Center
  City

16
Rule of Thumb

• Avoid generating choropleth maps of crime rates
  for small areas like city blocks because spurious
  results could be produced for areas where there
  are crimes but no residents.

17
(No Transcript)
18
Percent of Total Crimes
19
(No Transcript)
20
(No Transcript)
21
3d Maps???
22
Unique Grid Maps
23
Raster Maps
24
Raster Maps
25
Raster Maps
26
Raster Maps
27
Thematic Maps Linear Symbols

• Usually representing street features
• Links between where vehicles were stolen, where
  they were recovered, and suspects addresses
• Routes between victims and offenders home
  addresses
• Passes along streets by patrol cars to illustrate
  patrol density
• Traffic Volume

28
Classification In Maps

• How to classify data and represent symbols?
• Choice of scale controls abstraction
• Small scale maps with local detail ???????
• Too much detail becomes confusing and noisy
• Too little detail does not show anything
• Therefore there needs to be a balanced based on
  the message YOU want to convey
• JUST LIKE WRITING A MEMO

29
Classification In Maps

• Choropleth shading of crimes per area???
• Distortion due to area size
• Choropleth shading of crimes per 1000 persons
• Less area distortion ..
• Two categories ???
• Twenty categories???
• General rule Use no more than sixand not less
  than fourclasses, or shading levels

30
Classification In Maps

• Now we have the number of classes .. Then how to
  break them down.
• Equal ranges or intervals
• How is this done ?
• Equal Counts (Quantiles)
• Deciles? Quintiles? Quartiles?

31
Classification In Maps

• Natural Breaks finds low frequency points in
  the distribution.
• Jenks Optimization that ensures the internal
  homogeneity within classes while maintaining the
  heterogeneity among the classes
• Standard deviation (SD)
• 2 sd above the mean ??? What would this mean?
• 2 sd below the mean ??? What would this mean?
• What is the possible problem with the mean?

32
Classification In Maps

• Customized or manual often chosen due to a
  specific message to be sent

33
Classification In Maps
34
Exploratory Data Analysis

• First understand what the data distribution looks
  like
• Cases of Drunkeness by city block
• Normally distributed ???
• Would the mean value make sense?
• What is SKEWNESS?

35
Exploratory Data Analysis

• A normal distribution is the familiar bell-shaped
  curve that is seldom seen in crime data.
• Most crime data are positively skewed, meaning
  there is a long right "tail" representing a few
  high values.
• Hot spots (high crime areas) are geographic
  expressions of skewness, which presents
  difficulties in mapping numerical data.

36
Exploratory Data Analysis

• Equal Counts as a method????
• Standard Deviations as a method???
• Natural breaks or equal intervals will be the
  best methods for creating area-type maps in crime
  analysis. WHY?

37

• Census data