MAPPING AND ANALYTICAL APPLICATIONS OF GIS IN CRIME AND SECURITY ANALYSIS

1
MAPPING AND ANALYTICAL APPLICATIONS OF GIS IN
CRIME AND SECURITY ANALYSIS by Dr. Yuri
Gorokhovich, Assistant Professor, SUNY Purchase
College Dr. Victor Goldsmith, Associate Provost
for Research and Economic Development, Pace
University
First NSF/NIJ Symposium on Intelligence and
Security Informatics (ISI 2003) Tucson, AZ June
2-3, 2003
2
Context of Research Project Safe
Neighborhood Project Safe Neighborhoods (PSN) is
a nationwide commitment to reduce gun crime in
America by networking existing local programs
that target gun crime and providing those
programs with additional tools necessary to be
successful. The Bush Administration will seek to
commit 533 million to this effort over two
years, including 233.6 million already available
for this year. This funding will be used to hire
new federal and state prosecutors, support
investigators, provide training, and develop and
promote community outreach efforts.
3

• The Purpose of the Talk
• Discuss various GIS mapping and analytical
  techniques.
• Show their application to ATF data.
• Demonstrate complex relationship between data
  variables
• Demonstrate importance of combination of various
  GIS techniques to create better hot spots.

4

• PSN Program Goals
• The goals of this program are
• to increase the capacity of PSN Task Forces to
  design data driven strategies that produce
  measurable decreases in firearm-related violent
  crime
• to improve the long term ability of federal,
  state, and local partners to work together to
  understand, prosecute, and prevent firearm
  related violent crime within their jurisdictions.

5
NEW YORK STATE JUDICIAL DISTRICTS OF THE STUDY
EASTERN DISTRICT Suffolk Nassau Queens Brookly
n SOUTHERN DISTRICT Westchester Bronx Manhattan
Staten Island
6
ATF Gun Recoveries 1999 - 2003
District County Number of Gun Recoveries
Southern Southern Southern
Bronx 6696
Manhattan 743
Staten Island 979
Westchester 187
Eastern Eastern Eastern
Brooklyn 11330
Queens 2631
Nassau 913
Suffolk 901
7
Variety of Gun Dealers Premises per District and
County (top five)
District County Dealers Premise State Number of recoveries
Southern Bronx FL OH CA VA NY 1305 1037 1010 949 881
Manhattan OH CA FL CT SC 173 161 158 118 115
Staten Island NY FL OH CA CT 231 213 171 151 132
Westchester NY FL VA OH SC 50 32 22 22 20
8
Variety of Gun Dealers Premises per District and
County (top five)
District County Dealers Premise State Number of recoveries
Eastern Brooklyn FL CA OH NY MA 2076 1726 1664 1479 1274
Queens NY FL CA CT OH 501 468 348 315 308
Suffolk NY FL CT CA MA 258 169 169 137 136
Nassau FL NY CA MA OH 186 154 137 128 120
9
Variety of Gun Types per District and County
District County Gun Type Number of recoveries
Southern Bronx P PR S R PD Z M DD 4088 1670 433 401 88 8 5 3
Manhattan P PR R S PD 494 173 33 31 12
Staten Island P PR R S PD Z 483 230 177 81 7 1
Westchester P PR S R PD M 84 56 21 17 8 1
Codes P - pistol PR - revolver S - shotgun R -
rifle PD - derringer Z - unknown M - machine
gun DD - destructive device
10
Variety of Gun Types per District and County
District County Gun Type Number of recoveries
Eastern Brooklyn P PR R S PD Z M C DD 6521 3123 807 706 139 16 12 5 1
Queens P PR R S PD Z M 1414 612 301 257 41 4 2
Suffolk P PR R S PD Z M C 400 210 166 104 17 2 1 1
Nassau P PR R S PD Z 479 277 83 61 12 1
Codes P - pistol PR - revolver S - shotgun R -
rifle PD - derringer Z - unknown M - machine
gun DD - destructive device
11
Variety of Gun Recovery Locations (top five) per
District and County
District County Recovery location Number of recoveries
Southern Bronx Unknown 1 24 17 19 6194 322 99 46 7
Manhattan Unknown 1 24 17 12 622 43 29 23 8
Staten Island Unknown 1 24 17 16 908 28 27 4 3
Westchester Unknown 1 22 16 13 181 3 1 1 1
Code 1 intersection 7 police station 12 -
vehicle 13 airport 16 - bar 17 apartment 19
store 22 vacant lot 23 business 24 - residence
12
Variety of Gun Recovery Locations (top five) per
District and County
District County Recovery Location Number of recoveries
Eastern Brooklyn Unknown 1 24 17 19 10444 605 123 82 22
Queens Unknown 1 17 24 13 2352 142 46 45 21
Suffolk Unknown 24 23 22 19 824 60 7 3 2
Nassau Unknown 24 7 1 17 808 34 29 21 7
Code 1 intersection 7 police station 12 -
vehicle 13 airport 16 - bar 17 apartment 19
store 22 vacant lot 23 business 24 - residence
13

• Complexities of ATFE Database
• Gun dealers come from various states outside NY
  districts and some gun dealerships supply guns
  from NY
• Most of recovery locations are not known. From
  available data the majority of recovery locations
  are associated with intersections residences
  apartments bars stores vacant lots airports
• There is apparent relationship with locations of
  existing crime patterns (hot spots), schools and
  other public facilities and land uses.

14

• SPATIAL UNITS IN CRIME ANALYSIS
• There are three kind of spatial units stored in
  ATFE database that can be used for crime mapping
  
• Geographic coordinates (point), derived from
  street address
• Zip code (polygon)
• Administrative unit i.e. borough, town, state
  (polygon)
• All spatial units correspond to one of the
  following topological sequence
• Point Line Polygon
• In this sequence, a point represents lowest level
  of topology and serves as an elementary building
  block for all other topological types (i.e. lines
  and polygons)

15
SPATIAL UNITS IN CRIME ANALYSIS ACCURACY The
most accurate unit is geographic coordinate pair.
It allows pinpoint crime spot and can be used
in variety of mapping techniques, including
density mapping and proximity analysis among
others. Zip code and administrative boundaries
are relatively large areas and have high spatial
uncertainty within their boundaries. CONTEXT
Administrative boundaries (including boroughs,
villages, townships) have their own regulations
on traffic, parking, general security and land
uses. These factors can affect crime and gun
possession. From this point of view these
administrative boundaries can be used as spatial
units for analysis. While spatial techniques such
as density mapping or proximity analysis are
difficult to implement with these units, they
would be suitable for the multivariate or
descriptive statistics.
16
SPATIAL UNITS IN CRIME ANALYSIS CONCLUSION Geog
raphic locations (i.e. points) of recovery data
are more useful for the analysis of crime because
they represent lowest level of topology in
spatial data, provide basis for other topologic
types and can be analyzed with any GIS technique.
17
Address Matching and Geocoding Technique Using
Census TIGER Street Data
18
Address Matching and Geocoding (Interpolation
Using TIGER Data)
Fulton St
X1, Y1
Coordinate Pair Longitude Latitude
Length
From address right 135
X2, Y2
To address right 167
150 Fulton St
19
CRIME MAPPING TECHNIQUES THEMATIC MAPS NOT
REPRESENTATIVE, HIGH SPATIAL UNCERTAINTY
20
CRIME MAPPING USING POINT DATA ATTRIBUTES
Recovery of guns in Staten Island and Brooklyn,
1999 - 2003
Red dots are for 23,315 recoveries for the time
period
Data Source ATFE, West Virginia
21
CRIME MAPPING USING POINT DATA ATTRIBUTES
Recovery of guns in Staten Island and Brooklyn,
1999 - 2003
Red dots indicate locations with gt 5 recoveries
Data Source ATFE, West Virginia
22
CRIME MAPPING USING POINT DATA ATTRIBUTES
Recovery of guns in Staten Island and Brooklyn,
1999 - 2003
Red dots indicate locations with gt 20 recoveries
Data Source ATFE, West Virginia
23
CRIME MAPPING USING POINT DATA ATTRIBUTES
Recovery of guns in Staten Island and Brooklyn,
1999 - 2003
Red dots indicate locations with gt 50 recoveries
Data Source ATFE, West Virginia
24
CRIME MAPPING TECHNIQUES Unique Symbol
Recovery of guns in Bronx, 1999 - 2003
Typical Point Data Map
25
CRIME MAPPING TECHNIQUES Graduated Symbol
Color Codes
Gun Recoveries Number of gun dealers associated
with the recovery location
Legend Number of dealers per recovery
location
Data Source ATFE, West Virginia
26
CRIME MAPPING TECHNIQUES Graduated Symbol
Color Codes
Gun Recoveries Number of gun types associated
with the recovery location
Legend Number of gun types per recovery
location
1

-

2


3

-

4


5

-

6
Data Source ATFE, West Virginia
27
CRIME MAPPING TECHNIQUES Proximity of Recovery
Locations
Red segments represent lt 0.5 miles proximity
between recovery locations
Gun neighborhoods
Data Source ATFE, West Virginia
28
CRIME MAPPING TECHNIQUES Contour Line Densities
Denser Pattern of Contour Lines Corresponds to
the Higher Level of Recoveries
Bronx
Advantage of Contour Line Densities They show
orientation of the recovery field i.e. changes
as a result of police intervention strategies.
Data Source ATFE, West Virginia
29
CRIME MAPPING TECHNIQUES
HOT SPOTS TRADITIONAL DENSITY ANALYSIS
Recovery of guns in Staten Island and Brooklyn,
1999 - 2003
Number of recoveries per sq. mile
Data Source ATFE, West Virginia
30
CRIME MAPPING TECHNIQUES 3D Visualization
Points are recovery locations. Intensity of color
shows rate of recovery.
Data Source ATFE, West Virginia
31
CRIME ANALYSIS USE OF DEMOGRAPHIC SPATIAL DATA
Census Tracts
Census tracts allow the extraction of demographic
data and to overlay demographics with crime data.
The resultant mapping and analysis will lack
spatial pattern and precision as previously shown.
32
CRIME ANALYSIS USE OF DEMOGRAPHIC SPATIAL DATA
Census Blocks
Census Blocks allow to extract demographic data
and overlay them with crime data. Their spatial
pattern and precision are finer than census
tracts.
33
CONCLUSIONS

1. Density mapping, proximity and demographic
   analysis are most important categories of
   scientific tools for research on crime.
   Geographic coordinates provide basis for these
   analyses. Therefore their collection is very
   important. However, for locating crime and report
   preparation street addresses are traditionally
   used by police and other municipal agencies. GIS
   geocoding technique links street addresses and
   geographic locations if street data contain
   correct street address and zip code/city/borough.
2. Available GIS analytical and mapping techniques
   allow to create new kind of hot spots, based on
   combination of multiple factors from ATFE
   database, such as frequency of gun recoveries,
   frequency of recovery dates and proximity between
   recovery locations. Traditional hot spot
   technique is based primarily on recovery
   frequencies and uses an arbitrary radius for
   the kernel smoothing technique.
3. Design of GPS coupled with GIS-based street map
   would allow automatic collection of crime data in
   the field in both formats street address and
   geographic location. This would save time for
   police officer preparing report, reduce cost of
   initial data preparation for analysis and allow
   use of data in real-time response measure if
   needed.