i need someone to write me a summary with out plagiarism

This article was downloaded by: [Texas A&M University Corpus Christi] On: 13 April 2015, At: 15:54 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Annals of GIS Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tagi20

Why police and policing need GIS: an overview Fahui Wang a a Department of Geography & Anthropology , Louisiana State University , Baton Rouge , LA , USA Published online: 05 Jul 2012.

To cite this article: Fahui Wang (2012) Why police and policing need GIS: an overview, Annals of GIS, 18:3, 159-171, DOI: 10.1080/19475683.2012.691900

To link to this article: http://dx.doi.org/10.1080/19475683.2012.691900

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Annals of GIS Vol. 18, No. 3, September 2012, 159–171

Why police and policing need GIS: an overview

Fahui Wang*

Department of Geography & Anthropology, Louisiana State University, Baton Rouge, LA, USA

(Received 16 February 2012; final version received 19 April 2012)

This article provides an overview of major application areas of geographic information systems (GISs) in police departments and policing practices, mainly in the United States. It begins with a quick review of historical evolution of GIS applications in law enforcement from prototype computerized crime mapping to sophisticated spatial analysis and its integration in policing practices. Six major areas of GIS applications in police and policing are discussed with case studies to illustrate the value of GIS as (1) a partner for field officers, (2) a crime investigation and prevention tool, (3) a policy implementation and evaluation tool, (4) a police force planning tool, (5) a tool for testing crime theories, and (6) a communication tool. The article is concluded with an outlook for some emerging trends.

Keywords: GIS; crime analysis; law enforcement; policing; review

1. Evolution of geographic information system in crime analysis and law enforcement

As in many application areas, mapping is the foremost function of geographic information system (GIS) in crime analysis. The review begins with tracing the concept and practice of crime mapping. As early as in 1829, the Italian geographer Adriano Balbi and the French lawyer Andre- Michel Guerry used data from the first French national system of crime reporting to draw three choropleth maps of crimes (crimes against persons, crimes against prop- erty, and school instruction) in France on a large sheet of paper (as reviewed by Kenwitz 1987). The maps revealed how the geographic patterns of various crimes were related to each other and also to socioeconomic indicators such as educational levels. The ecological perspective of crime was expanded by Lambert Adolphe Quetelet, a Belgian astronomer-turned social statistician, who linked crime to more contextual variables such as climate, poverty, trans- portation routes, and ethnic and cultural variations in addi- tion to educational levels (as reported in Stigler 1997). In the United States, mapping of crime and related social variables was widely used by the urban ecologists at the University of Chicago (e.g., Park 1915, Sutherland 1924, Shaw and McKay 1942). This group of sociologists rep- resenting the Chicago School was credited for developing the social disorganization theory for crime. The research of ecological correlates of crime such as juvenile delinquency ‘necessarily begins with a study of its geographical loca- tion’ (Shaw 1929, p. 5), and thus naturally relies on maps.

*Email: [email protected]

However, such an interest quickly faded, and attention was shifted away from geographic variation and oriented to the impacts of social disorganization on individual behaviors (Robinson 1982). Such a shift may be partially attributable to the large amount of time and effort required for col- lecting crime data and the arduous process of manually drawing maps (Weisburd and McEwen 1997, p. 7). In the meantime, ‘spot maps’ were a common practice in police departments in the United States by physically placing pins on large street maps to indicate crime hazards (FBI 1944). It seemed evident that both academic researchers and practitioners could benefit from automated crime mapping.

Among early experiments of computerized crime map- ping, Pauly et al. (1967) used the SYMAP program devel- oped at Harvard University to map the distribution of crimes in St. Louis, Missouri, at about the same time the first GIS (i.e., Canada GIS) was developed (Clarke 1999, p. 9). At the time, computerized crime mapping required major hardware investment and programming efforts, and technologies for computer-based crime map- ping were primitive. For example, shading of a black-and- white choropleth map was implemented by overstriking printer characters for darker symbols. The quality of maps was understandably subpar, and yet the technique generated much excitement, and was strongly advocated by several police organizations (Buch et al. 1973, Chang et al. 1979). The major draw is that the maps helped police focus more resources on areas of high crime frequencies. As desk- top computers became more affordable along with more

ISSN 1947-5683 print/ISSN 1947-5691 online © 2012 Taylor & Francis http://dx.doi.org/10.1080/19475683.2012.691900 http://www.tandfonline.com

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

160 F. Wang

user-friendly GIS software, GIS development passed the stages of ‘pioneer age’ from the mid-1950 to the 1970s and ‘research and development age’ and ‘implementation and vendor age’ in the 1980s, and entered the stage of ‘client applications age’ in the 1990s (Foresman 1998). Its applications in law enforcement experienced a signif- icant growth in the late 1980s and 1990s. Based on a 1997 survey conducted by the Crime Mapping Research Center (CMRC) of the National Institute of Justice (NIJ, the research arm of the US Department of Justice), comput- erized crime mapping in law enforcement agencies in the United States has experienced rapid growth since the late 1980s, as shown in Figure 1 (Mamalian et al. 1999). The trend has continued into the twenty-first century (Weisburd and Lum 2005) and resembles the S-curve of innovation diffusion (Rogers 1995).

Three forces came together to make the ‘perfect storm’ for the rapid growth of GIS adoption in law enforcement agencies and in crime research since the late 1980s. The first and foremost was the advancement of computer and information technologies. Expensive mainframe comput- ers were replaced by cheaper and more efficient desktop computers, and developments in mapping and GIS software made them more powerful and user friendly. Second, crime data reporting and archiving became mandatory in the US criminal justice system. In particular, accurate location information (e.g., street address) for crime incidences has been required in modern police departments. In the mean- time, other supporting data in GIS format also became standardized and available. For instance, the obscure Dual Independent Map Encoding spatial data model used by the US census till 1980 was replaced by the more accessi- ble Topologically Integrated Geocoding and Referencing in 1990. The third factor was the theoretical shift in criminology from ‘offender-based approaches’ to a focus on ‘context of crime,’ that is, the ‘physical, organiza- tional, and social environments that make crime possible’

1997 CMRC survey: cumulative distribution of crime mapping adoption

0 20 40 60 80

100 120 140 160 180 200

T ot

al n

u m

b er

o f

d ep

ar tm

en ts

19 82

?

19 83

?

19 84

?

19 85

?

19 86

?

19 87

?

19 88

?

19 89

?

19 90

?

19 91

?

19 92

?

19 93

?

19 94

?

19 95

?

19 96

?

19 97

?

Year

Figure 1. Growth of computerized crime mapping in law enforcement agencies in the United States (Mamalian et al. 1999).

(Weisburd and McEwen 1997, p. 14). Among various the- ories of crime, the routine activities approach developed by Cohen and Felson (1979) explains the occurrence of crime as the confluence of three elements: a motivated offender, a desirable target, and the absence of a capable guardian. This perspective emphasizes the interaction of the three elements at the same place and at the same time for a crime to occur and thus has a focus on ‘place.’ Similarly, the situational crime prevention approach pioneered by Clarke (1980) suggests reducing criminal opportunities in very specific contexts as an effective crime control strat- egy. The rational choice theory (Cornish and Clarke 1986) argues that a potential offender consciously evaluates the expected costs and rewards of crime versus alternative behaviors and, thus, provides the theoretical foundation for the situational crime prevention measures. Environmental criminology advocated by Brantingham and Brantingham (1981) has an unequivocal focus on place. This theoretical shift has also helped change the fate of crime geographers from being ‘displaced’ or marginalized to a major source of knowledge and innovations for other criminological fields (LeBeau and Leitner 2011).

GIS and related technologies have turned crime map- ping to a powerful decision-making tool for law enforce- ment agencies. However, GIS is beyond mapping, and increasingly so for crime analysis and law enforcement (Wang 2005a). The NIJ has been instrumental in promot- ing the use of GIS and spatial analysis in crime research (Wilson 2007). The efforts are exemplified in fund- ing the development of CrimeStat (http://www.nedlevine. com/nedlevine17.htm) and organizing the annual Crime Mapping Research Conference (http://www.nij.gov/events/ maps/welcome.htm). Early applications of GIS in crime analysis were limited to data archiving, automated pin mapping, and cluster or hot-spot analysis (Harries 1999, p. 94), and gradually extended to applications of spatial statistics in testing spatial autocorrelation (e.g., Wasserman and Stack 1993, Messner et al. 1999) and construct- ing spatial-lag variables in regression (e.g., Roncek and Montgomery 1995, Baller et al. 2001, Morenoff et al. 2001). Most recently, collaborations among the academic (criminologists, geographers, computer scientists, statisti- cians, and others), business, and practitioner communities have led to important advancements, and GIS applications have evolved from crime mapping to sophisticated spatial analysis and its integration in daily policing practices, as elaborated in the next six sections on major application areas of GIS in police and policing.

2. GIS as a partner for field officers

The advancement of geospatial technologies has made GIS a trusted ‘partner’ for police officers. This section discusses how GIS is used in data reporting, mobile solutions, and protecting police from lawsuits.

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 161

The quality of data forms the foundation for creditable research as well as data-driven intelligence-led policing (Ratcliffe 2003). Crime underreporting is a common prob- lem at various levels of jurisdictions. Crime researchers often favor the use of data of serious crimes (e.g., homi- cide) more than data of other crimes because of gen- eral confidence in the quality of such data (Land et al. 1990). However, even the data of serious crimes are not immune to the problem of data uncertainty. In the United States, there have been significant discrepancies, though converging over time, in homicide data between the Federal Bureau of Investigation’s (FBI) Uniform Crime Reporting (UCR) Program focusing on offenders and the Bureau of Justice Statistics’ (BJS) National Crime Victimization Survey (NCVS) from the victim’s perspec- tive (Ansari 2010). According to a report released by the Home Office, the United Kingdom experienced a sudden 22% yearly rise in the most serious violent crimes (mur- der, manslaughter, and severe assaults) in 2008 because of underreporting of these crimes in the prior years (Swaine 2008). Crimes could also be overreported, par- ticularly in criminal victimization surveys (Levine 1976). Inaccuracy in crime data invites the criticism of ‘garbage in and garbage out’ and undermines the validity of crime studies.

For spatially based studies and policing practices, geocoding accuracy is a major issue for compiling crime data. Traditional geocoding is based on address. Sources for address-based geocoding errors include typographic errors, abbreviations, address duplication, lack of standard- ization, etc. (Ratcliffe 2004). Errors may also be caused by loss of concentration by officers in the tedious address- typing process, evidenced in a trend of increasing errors in data input toward the end of the night shift (McCarthy and Ratcliffe 2005). Modern GIS systems installed on patrol cars or handheld GIS devices (e.g., Cody Mobile 7, Spillman’s Mobile Office) detect and correct address errors and increasingly rely on global positioning system (GPS) for geocoding to release field officers from much of the data entry labor. Automated law enforcement field inter- view forms are also used to minimize other errors and save officers time in crime reporting. A fully integrated GIS- based mobile office for field officers does more than data reporting. One useful function is to link to other contex- tual information (neighborhood demography, business and natural environments, historical crime records, etc.) so that officers on patrol or dispatched in response to a call can be better prepared on assignments. An automatic vehicle loca- tion (AVL) module tracks the location of all GPS-equipped fleet units in real time, identifies the closest officer for backup, or optimizes response times by dispatching the unit closest to a call by accounting for street network and barriers. Information from the AVL system is fed back to the command center to help improve the performance of computer-aided dispatch services. A license plate reader

module gathers information of vehicles by a camera and helps locate stolen or wanted vehicles.

Maps, modernized by GPS and GIS technologies, have been used as a powerful tool in lawsuit. Here, a case study illustrates how GIS was used to defend an officer (Smith 2011). In 1999, a felon after being arrested and con- victed filed a lawsuit against the City of Shawnee, Kansas, claiming that police used excessive force while making the arrest. A map (shown in Figure 2) was made in GIS to assist the arresting officer’s testimony. It showed that the suspect travelled at high speeds through residential areas, running several red lights and bailing out of his car in a densely populated housing project in order to evade the police. The offender was considered posing a great danger to public safety, and the officer was cleared from the case.

3. GIS as a crime investigation and prevention tool

This section discusses how GIS is used as an effective tool in crime investigation and prevention. There are ample of cases that have demonstrated the value of GIS in solving crimes. Many murder cases were solved by police track- ing cell phones used by victims (e.g., see the report on solving the murder case of Kelsey Smith at http://www. kmbc.com/r/13453443/detail.html), or tracking suspects by GPS (e.g., see the report on the Brad Jackson case in Leipnik et al. 2003). One caution needs to be taken since the use of GPS for surveillance and evidence gathering may involve a legal issue related to protection of privacy in the United States, exemplified in the most recent Supreme Court’s decision on the US versus Jones case (http:// www.supremecourt.gov/opinions/11pdf/10-1259.pdf). The following discussion focuses on geographic profiling and predictive policing because both use more advanced spatial analysis.

Geographic profiling has received much media cover- age after it helped solve several high-profile crimes. The technique is to predict an unknown criminal’s most proba- ble residence area by analyzing the geographic locations of a linked series of crimes. Its theoretical foundation is the well-established finding of distance decay behav- ior on a criminal’s journey to crime (J2C) from his/her residence to crime sites (Van Koppen and De Keijser 1997, Rengert et al. 1999). In other words, an offender commits most offenses fairly close to home, and the pro- portion of offenses committed decreases with increasing distance. In addition, the criminal also tries to avoid com- mitting crimes too close to home, within a mental ‘buffer zone’ (Brantingham and Brantingham 1981), as shown in Figure 3. Based on a distance decay function for J2C, a probability surface is generated for every point in a study area to represent the likelihood that any given point is the offender’s base (most commonly home or work site). Various functions and parameters have been proposed, most available in CrimeStat, to model the distance decay

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

162 F. Wang

Figure 2. Paths of suspect and officer and surrounding areas in a legal case (Smith 2011).

behavior (Levine and Block 2011). Some suggest more fac- tors than distance influencing the J2C trips (e.g., Rattner and Portnov 2007), and others argue that the choices of function forms have minimal effects on the outcome of geographic profiling (Hammond and Youngs 2011). ‘Distance’ can also be measured as Euclidean, Manhattan (as in Rossomo et al. 2005), shortest travel path, or quickest temporal path (as in Kent et al. 2006). Research indicates that not a single function or a distance measure is favored because of the complexity of an offender’s behavior that is ‘predicated on the individual’s mental map’ – one’s own perception of the environment (Kent et al. 2006, p. 197). Geographic profiling is usually applied to the analysis of series/repeated crimes because data volume is critical for its success. It is intended to be used, in combination with other information and investigative techniques, to prioritize leads or direct the geographic focus of an investigation.

In addition to crime investigation, GIS plays a promi- nent role in predictive policing, which uses information and advanced analysis to inform forward-thinking crime prevention (Uchida 2010). By identifying environmental

Probability of target selection

Buffer zone

Residence Distance

Figure 3. Crime distance-decay function with buffer zone (Rossomo et al. 2005, p. 104).

factors associated with past events, analysts can predict areas statistically similar to locations where prior incidents occurred. In contrast to hot-spot policing, which focuses on areas with a high density of past events (discussed in Section 5), this approach also detects emerging areas with changed environmental factors and thus allows law

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 163

enforcement to ‘proactively prevent and disrupt crime’ in targeted neighborhoods (McCue 2011, p. 4). Gorr and McKay (2005) applied a ‘tracking signal technique’ widely used in management science to estimate trend and season- ality and quickly detect step jumps and outliers in crime trends. The method filters out regular noise and focuses on detecting departures from established time trend pat- terns. Based on the method with data inputs such as 5-year monthly crime counts and land use, a crime early warn- ing system (CEWS) was developed for 1-month-ahead crime forecasts across 100 grid-cell areas in Pittsburgh. The system helps police stay one step ahead of criminals and proves to be a promising tool in proactive policing. Predictive policing can be also successful in preventing individual crimes. A recent example is the case of the ‘blue bandana bandit’ in Glendale, Arizona. Based on a series of robberies, crime analysts used GIS and CrimeStat to predict where the next one might take place, and police suc- cessfully made an arrest at the predicted site (http://www. nij.gov/journals/253/predicting.html).

4. GIS as a policy implementation and evaluation tool

Some law enforcement and correction tasks have an explicit spatial component that requires the use of GIS- based technologies for monitoring and evaluation.

On 20 May 2011, the former IMF chief Dominique Strauss-Kahn was remanded to home confinement in a New York apartment, and his whereabouts were moni- tored by a GPS device (http://www.csmonitor.com/USA/ 2011/0520/House-arrest-What-ex-IMF-chief-Dominique- Strauss-Kahn-can-expect). High-profile legal cases such as this have certainly helped educate the public on the value of GPS tracking in law enforcement. However, the use of GPS and GIS is not limited to monitoring celebrities. For example, the Electronic Monitoring Protection and Crime Tracking (EMPACT) project in Florida automatically correlated data from GPS tracking of offenders (proba- tioners, parolees, and offenders on pretrial release) and local crime incident data through a web-based interface, and determined whether a tracked offender was at the scene of a crime incident (Frost 2005). In the United States since 2003, Megan’s Law has placed spatial restrictions on the residential locations of sex offenders, prohibiting them from residing within 1000 feet of any school-related properties (Levenson and Cotter 2005). The implemen- tation of Megan’s Law relies on the proximity analysis tool (e.g., buffer) available in any major commercial GIS software. A study in Hamilton County of Ohio assessed the enforcement of the law and revealed that 45% of sex offenders lived within the 1000-foot school restriction zone (Grubesic et al. 2007), raising other social issues such as housing availability and equity.

One issue in evaluating the effectiveness of crime control strategies is spatial externalities associated with

these activities, whose measures could benefit from GIS- supported spatial analysis. There are two types of spatial externality. Spatial displacement happens when crime con- trol measures cause crime to move away, and therefore, crimes are reduced in target areas but increased elsewhere. In this instance, there is no overall reduction in crime. Spatial diffusion, in contrast, means that the benefits of crime reduction ‘spill over’ to neighbors (free riders) and lead to a decline of crime in nearby areas as well and is thus more desirable. Both are unintended side effects or spatial externalities, and lead to crime rates in nearby areas to correlate with each other, that is, spatial autocor- relation. Displacement is derived from the rational choice theory (as discussed in Section 1). If some crime preven- tion action is taken in one area, those seeking to gain from crime move to other areas where detection and apprehen- sion risks are expected to be lower (Cornish and Clarke 1987). Most criminals also obey the distance decay rule (as discussed in Section 3) or the familiarity decay rule (Eck 1993), and take a short journey to commit crimes in nearby areas. Therefore, in a spatial context, a decline in the crime rate in one area is associated with an increase in the crime rate in its nearby areas, and vice versa. Crime diffusion or free-rider effect can be also explained by the rational choice theory. As criminals have imperfect information, some believe that certain areas have enhanced protection but do not know the exact areas. Consequently, they are deterred from sites with protection as well as sites with- out protection (Miethe 1991, Clarke and Weisburd 1994, Ayres and Levitt 1998). More likely nearby areas bene- fit first. Thus, crime decline in one area leads to crime reduction in adjacent areas as well. Several reviews (Eck 1993, Hesseling 1995, Braga 2001) found that spatial dis- placement cases were in the minority, and other argued that crime displacement in the short term coexisted with crime diffusion in the long run (Fabrikant 1979).

The weighted displacement quotient (WDQ), devel- oped by Bowers and Johnson (2003), has been used in several recent GIS-based empirical studies for evaluating the spatial externalities of crime prevention operations. The implementation of WDQ requires the definitions of three areas: (1) the target area where the crime reduction strategy has been applied, (2) a buffer area where crime would be displaced or the crime control benefits would be diffused, and (3) a control area that acts as a baseline on general crime trends in the region in general. The WDQ is written as

WDQ = Bt1 /

Ct1 − Bt0 /

Ct0 At1

/ Ct1 − At0

/ Ct0

where A, B, and C are crime counts in the target area, buffer area, and control area, respectively, and t0 and t1 are the time periods before and during the intervention.

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

164 F. Wang

0–1 1

Displacement increases Diffusion increases

WDQ

D if

fu si

on >

D ir

ec t

ef fe

ct s

D if

fu si

on =

D ir

ec t

ef fe

ct s

D if

fu si

on <

D ir

ec t

ef fe

ct s

N o

D if

fu si

on o

r D

is pl

ac em

en t

D is

pl ac

em en

t = D

ir ec

t ef

fe ct

s

D is

pl ac

em en

t < D

ir ec

t ef

fe ct

s

D is

pl ac

em en

t > D

ir ec

t ef

fe ct

s

Figure 4. Seven scenarios for WDQ-based evaluation of spatial externalities in policing.

The numerator (Bt1 /

Ct1 − Bt0 /

Ct0 ) represents a buffer displacement measure, and the denominator (At1

/ Ct1 −

At0 /

Ct0 ) represents a success measure. Based on Bowers and Johnson (2003), there are seven likely outcomes corre- sponding to the value of WDQ falling in different ranges as shown in Figure 4.

Hall and Liu (2009) used a modified spatiotemporal WDQ index to examine the effects of an aggressive polic- ing operation in Cincinnati, Ohio. Specifically, the control area was defined as the same area as the target area (or the buffer area) but with data from a prior time period in order to mitigate the problem of the control area being cor- rupted by large-scale displacement. By using various buffer sizes to define the buffer area, they found a diffusion of benefits to be taking place at the immediate neighboring areas, but crime displacement at a greater distance from the operation deployment area. In addition to examining the changes in crime density maps and based on the mean centers of crimes, Lowry and La Vigne (2011) also used the WDQ index to measure whether the installation of a camera surveillance system in a neighborhood in Chicago gener- ated spatial externalities in the crime pattern. The results revealed no definitive signs of diffusion of benefits beyond the camera viewsheds and no findings of displacement. Ratcliffe and Breen (2011) used the WDQ index to assess the effects from a series of directed and geographically tar- geted enforcement initiatives in Camden, New Jersey, in 2005. To address a major concern of the WDQ for lack of a measure of statistical significance, a phi statistic was devel- oped to measure the level of association between the target and the buffer areas. The study found a diffusion of benefits approximately equal to the direct effects of the operation in the target area.

5. GIS as a police force planning tool

The rising role of GIS in law enforcement in general and police force planning and deployment in particular can be

attributable to the turning tide of crime theory to focus on place instead of people, as explained in Section 1. The concentration of crime in limited areas is astonishing, and naturally invites place-based policing (Weisburd 2008). For example, 3% of the city’s addresses accounted for 50% of calls for service to the police in Minneapolis (Sherman et al. 1989), and about 4% of streets and intersection areas generated nearly 50% of the city’s narcotics arrests and almost 42% of the disorder arrests in Jersey City, New Jersey (Weisburd and Mazerolle 2000). This section dis- cusses two practices that rely on GIS for effective police force planning: hot-spot policing and police districting.

Hot-spot policing plans and adjusts the deployment of police force in accordance with the geographic varia- tion of crime and focuses police patrol on crime hot spots (Weisdud 2005). A common definition of hot spot is an area with ‘a greater than average number of criminal or disorder events’ (Eck et al. 2005, p. 8). Earlier studies of hot spots used basic methods such as visual interpre- tation, choropleth mapping, and grid-cell density analysis (Jefferis 1999). Grid-cell density analysis uses the kernel density estimation, the empirical Bayes estimation, and other spatial smoothing methods to convert discrete points of crime events to a continuous density map, which shows relative concentrations of crime in certain areas. However, local pockets of crime concentration or elevated crime rates may arise simply by chance alone, and require rig- orous statistical analysis to detect clusters (hot spots) that are statistically significant (Wang 2006, p. 167). Free soft- ware such as CrimeStat, SaTScan (www.satscan.org), or GeoDa (geodacenter.asu.edu) contain the tool that iden- tifies hot spots within regular shapes such as circles or eclipses. Attempts have also been made to detect flexi- bly shaped clusters (e.g., Grubesic and Murray 2001), as shown in Figure 5, but the challenge remains to develop robust geometric and statistical measures for determining cluster significance. Among the various spatial autocorre- lation indices, the G∗i statistic developed by Getis and Ord (1992) is used most often to identify local clusters with sta- tistically significant high or low crime rates (corresponding to hot and cold spots), available in the popular GIS software ArcGIS. Empirical studies have shown that hot-spot polic- ing is largely effective with no apparent or major crime displacement (Weisburd 2005).

The basic objectives in police districting include meet- ing the agency’s response time threshold, minimizing the cost of operation, and balancing workload across districts (Taylor and Huxley 1989). Few of the advanced mathe- matical models for districting, proposed several decades ago (Reinier et al. 1977), were implemented until the advancement of GIS and computer technology enabled the computation in a reasonable time period. Several case studies of police districting were presented in a recent issue of Geography & Public Safety (http://www.nij. gov/maps/gps-bulletin-v1i4.pdf). ArcGIS has a districting

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 165

Cluster 1 Cluster 2

Cluster 3

Hull/Member Ratio = 0.0662 Avg. Distance = 677.9

Hull/Member Ratio = 0.3904 Avg. Distance = 278.87

Hull/Member Ratio = 0.1554 Avg. Distance = 434.8

Figure 5. Detecting flexibly shaped clusters (Grubesic and Murray 2001, p. 9).

extension that helps analysts test various scenarios of district boundaries and review the outcomes including projected workloads (http://www.esri.com/software/arcgis/ extensions/districting/). However, studies integrating GIS and mathematical modeling for police districting remain a rarity. One exception is a study by Curtin et al. (2005) that applied an optimal covering model to define the most efficient police patrol areas (i.e., maximizing the number of incidents served while meeting an acceptable response time). The method took GIS data layers of incidents (e.g., service calls) and road network, used linear programming to formulate the optimization problem, and found heuristic solutions that increase the current level of police service. The study shows great potentials of ‘optimal districting.’ More work needs to be done to explore other goals such as minimizing total costs (response time), minimizing the number of districts (dispatch centers), maximizing equal accessibility or a combination of multiple goals (Wang 2012), and evaluate the benefits of various scenarios.

6. GIS as a tool for testing crime theories

There is a large body of literature on GIS applications in crime research. This section briefly reviews a few case studies of using GIS to test crime theories, all with

significant implications in methodological development in spatial analysis.

The first set of case studies uses the agent-based modeling (ABM) to simulate and test the routine activity theory. Liu et al. (2005) used a cellular automaton model to simulate street robbery patterns in a small neighborhood near downtown Cincinnati. Based on the routine activity theory, the model considered offenders, targets, and crime places as individual agents and simulated crime patterns based on the interaction between these three agents at a specific time. The model was implemented in Visual C++. The calibrated simulation model generated crimes simi- lar to actual crimes in both the total number of crimes and their spatial distribution, as shown in Figure 6. The work by Groff (2007a, 2007b, 2008) was similar in that ABM was used to examine robbery patterns based on the routine activity theory. In implementation, she used the software product Agent Analyst to pair the GIS software ArcGIS with the ABM software RepastPy (North et al. 2006). By doing so, individuals were modeled in a real geo- graphic environment and various scenarios were simulated through the inputs of ABM. Although the models reviewed here remain simple, the approach demonstrated its promise as a virtual laboratory for ‘exploring the impact of policy decisions’ (Groff 2008, p. 114), where various scenarios

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

166 F. Wang

1

2

5 8

7

9 13

14

19

10

18 3

4

15

16 1211

17

20

6

Historical crime pattern

(a) (b)

1 2–3 4–5 6–8 9–15 16–23

Streets

Simulated crime pattern

2–3 1

4–5 6–8 9–15

16–23

24–31 Streets

Figure 6. Spatial pattern of street robberies: (a) actual crimes 1997–98 and (b) cellular automaton-simulated crimes (Liu et al. 2005, p. 209).

of experiments are costly to design and implement. See Liu and Eck (2008) for more case studies of GIS-based simulation of crime.

The second set of case studies is related to the small population problem, often encountered in the analysis of crime data, where the rate estimates for rare crimes such as homicide in small populations are unstable and very sus- ceptible to data errors. One viable approach to mitigate the problem is to construct larger areas from small ones so that the base population is sufficiently large and compara- ble across areas, commonly referred to as ‘regionalization.’ Several GIS-based automated regionalization methods have been proposed to merge similar and adjacent areas to form larger regions including the AZP (Openshaw 1977, Grady and Enander 2009), MaxP (Duque et al. 2007), and REDCAP (Guo 2008). Wang and O’Brien (2005) used some regionalization methods to test the herding-culture-of-honor hypothesis proposed by Nisbett (1993) and Reaves (1992). The hypothesis argues that areas where topography or precipitation limited agriculture to marginal farming and herding (dry plain and moist hill counties) had a chronic threat of livestock theft and should have higher homicide rates than those areas where farming was the major form of agriculture (moist plain coun- ties). The result indicated that the herding-culture-of-honor hypothesis proposed by Nisbett (1993) and Reaves (1992) was merely an artifact of less reliable estimate of homi- cide rates in counties with small population than counties of large population in southeastern United States. In other studies, a regionalization method based on the scale-space theory was applied to the analysis of homicide rates in Chicago (Wang 2005b, Mu and Wang 2008). The results showed that the method was able to minimize informa- tion loss in data aggregation and mitigate the modifiable area unit problem, another problem common in analysis of geographic data.

The final set of case studies involves the construction of localized measures in crime studies. Wang and Minor (2002) used the GIS-based network analysis technique to measure a localized job accessibility index as the ratio of

jobs to resident workers within a reasonable commuting range from a residential location. The index value changed as the targeted residential location moved from place to place, and thus captured the variation of local job mar- ket condition across the study area. They found that the index was highly correlated with crime rates in Cleveland, Ohio. Another study by Wang and Arnold (2008) devel- oped a localized income inequality (LII) index to measure the relative inequality when people compare their situa- tion to those around them. The LII, defined as the ratio of average income of a neighborhood’s contiguous areas to its own income (Figure 7), added a new dimension to

j = 2 $30,000

j = 1 $70,000

i = 4 $25,000

i = 6 $60,000

Subject area

Contiguous area (Rook) Contiguous area (Queen)

i = 7 $50,000

i = 5 $40,000

j = 3 $50,000

Figure 7. Defining contiguous areas and localized income inequality (Wang and Arnold 2008, p. 263).

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 167

concentrated disadvantage in urban areas and made it a stronger predictor of homicide rates across various areal units in Chicago.

7. GIS as a communication tool

Two functions highlight the value of GIS as a commu- nication tool in law enforcement: (1) facilitating inter- jurisdictional and cross-departmental data sharing and (2) enhancing information dissemination to the public and encouraging public participation.

In the era of rapid growth in transportation and telecommunication, criminals are becoming increas- ingly mobile. The need for data (including spatial data) sharing among law enforcement agencies over differ- ent jurisdictions has become more pressing. Many law enforcement fusion centers have been created across the United States to integrate data from multiple agen- cies and help facilitate local, state, county, and federal data sharing. Even within the same municipality or county or state, lack of communication across various departments can seriously hinder the ability of data- led intelligent policing. GIS plays an important role in

breaking down barriers between jurisdictions and agencies. One example is the Community Mapping, Planning and Analysis for Safety Strategies (COMPASS) project in Milwaukee, Wisconsin (http://compass.milwaukee.gov/ website/compass/viewer.htm; Figure 8). The project brought together a wide range of data (traffic, property, crime, public services, health, etc.) from different depart- ments and used a web GIS platform to support better access to crime-relevant data and facilitate communi- cation among citizens, the scientific community, and policy-makers (Albrecht and Pingel 2005).

GIS is also used as an effective way to inform citi- zens of updated crime information in their neighborhood and empower citizens to take caution or report valuable information. For example, the Citizen Law Enforcement Analysis and Reporting (CLEAR) MAP web application, maintained by the Chicago Police Department (http://gis. chicagopolice.org/), enables citizens to search crimes in the past 90 days by an address, around a school or park, by var- ious geographic areas (community area, ward, police beat, and district). The ‘IdentifyLA’ (http://identifyla.lsu.edu) web site, maintained by the FACES laboratory of Louisiana State University, maps all missing and unidentified people

Figure 8. The COMPASS project in Milwaukee, Wisconsin.

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

168 F. Wang

in Louisiana online and allows users to search cases by keywords, location, parish, etc. The program invites wide citizen participation in solving pending cases. By doing so, GIS, particularly through various web GIS platforms, will become truly ‘people’s GIS,’ a concept promoted by those advocating participatory GIS (Dunn 2007).

8. Concluding remarks

The preceding sections provide a glimpse of major appli- cation areas of GIS in crime analysis and law enforcement. The overview is intended for readers, many working in the field of GIS but not familiar with its police applica- tions, or others in law enforcement but not aware of the full potentials of GIS. The related literature has grown tremendously along with the growth of GIS technologies. The topics and case studies (usually most relevant or most recent work) are chosen to illustrate related concepts. The limitation of space and the author’s paucity of knowledge and experience prohibit the reviewer from a more in-depth examination of the topics, each of which deserves a full paper to cover.

Here, concluding comments are directed to some areas with much work to do. One exciting emerging field is the use of GIS-related 3D visualization technologies to develop some avatar-like virtual world systems for police force training. The Federal Law Enforcement Training Center (FLETC) in the United States has a Simulation Integration Management Branch that is leading the efforts in this endeavor. Some of the existing police training cur- ricula already include the use of computer-based video games to train officers how to respond in simulated crime scenes. However, the quality of most of these training games remains primitive, and rare (but exerting major impacts) crimes such as those involving hostages or ter- rorist attacks are usually not covered. Several technology challenges need to be tackled in order to make it feasi- ble and affordable for popular use: capturing complex real world in 3D, efficiently managing a large volume of 3D data, designing a system interactive with planted crimi- nals, and modeling the senses of distance, blind side, shade and shadow, etc. Another area is to continue the pursuit of interdisciplinary collaborations to inject new perspec- tives, new methods, and new insights into spatially oriented crime analysis and policing. For example, the application of social network analysis (e.g., Radil et al. 2010) has shown promises in analyzing organized crimes. Implementing the complex network analysis in a spatial context takes an interdisciplinary team of criminologists, computer sci- entists, geographers, and perhaps physicists. Finally, the advancement of police GIS and the full cultivation of its benefits rely on mutual trust, effective communica- tion, and close collaboration between academia and law enforcement professionals. A practitioner may ‘perceive an academic’s research as suggestive’ (LeBeau and Leitner

2011, p. 167) and, thus, hinder its dissemination. One proven effective channel to bridge the two is ‘career migra- tion’ between the two. A practitioner-turned academic or an academic-turned practitioner certainly commands more respect from the colleagues working in the conjunction of GIS and policing.

Acknowledgements An earlier version of the article was presented as a keynote speech at the National Conference on GIS Applications in Police in Beijing, 19–20 July 2011. I benefited from another keynote speech by Susan Smith, and insightful comments by Rixing He and Daliang Zhou at the Conference.

References Albrecht, J. and Pingel, J., 2005. GIS as a communication pro-

cess: experiences from the Milwaukee COMPASS project. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 1–23.

Ansari, S., 2010. Understanding and modeling the conver- gence of the UCR and NCVS: a time series analysis [online]. Criminology and Justice Policy Dissertations. Paper 2. Available from: http://hdl.handle.net/2047/d20000635 [Accessed 21 May 2012].

Ayres, I. and Levitt, S.D., 1998. Measuring positive externali- ties for unobservable victim precautions. Quarterly Journal of Economics, 113, 43–78.

Baller, R.D., et al., 2001. Structural covariates of U.S. county homicide rates: incorporating spatial effects. Criminology, 39, 561–590.

Bowers, K.J. and Johnson, S.D., 2003. Measuring the geograph- ical displacement and diffusion of benefit effects of crime prevention activity. Journal of Quantitative Criminology, 19, 275–301.

Braga, A.A., 2001. The effects of hot spots policing on crime. The Annals of American Political and Social Science, 578, 104–125.

Brantingham, P.J. and Brantingham, P.L., 1981. Environmental criminology. Beverly Hills, CA: Sage Publications.

Buck, G.A., et al., 1973. Police crime analysis unit hand- book. Washington, DC: Law Enforcement Assistance Administration, US Department of Justice.

Chang, S.K., et al., 1979. Crime analysis system support: descrip- tive report of manual and automated crime analysis functions. Gaithersburg, MD: International Association of Chiefs of Police.

Clarke, K.C., 1999. Getting started with geographic information systems. Upper Saddle River, NJ: Prentice Hall.

Clarke, R., 1980. Situational crime prevention: theory and prac- tice. British Journal of Criminology, 20, 136–147.

Clarke, R.V. and Weisburd, D., 1994. Diffusion of crime control benefits: observations on the reverse of displacement. Crime Prevention Studies, 2, 165–184.

Cohen, L. and Felson, M., 1979. Social change and crime rate trends: a routine activity approach. American Sociological Review, 44, 588–605.

Cornish, D.B. and Clarke, R.V., 1986. The reasoning crimi- nal: rational choice perspectives on offending. New York: Springer-Verlag.

Cornish, D.B. and Clarke, R.V., 1987. Understanding crime displacement: an application of rational choice theory. Criminology, 25, 933–943.

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 169

Curtin, K.M., et al., 2005. Integrating GIS and maximal cov- ering models to determine optimal police patrol areas. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 214–235.

Dunn, C.E., 2007. Participatory GIS – a people’s GIS? Progress in Human Geography, 31, 616–637.

Duque, J.C., Anselin, L., and Rey, S.J., 2007. The max-p region problem: Regional Analysis Laboratory (http://regionalanalysislab.org/) [online]. Working Paper No. 20070301. Available from: http://www.pysal.org/users/ tutorials/region.html [Accessed 21 May 2012].

Eck, J.E., 1993. The threat of crime displacement. Criminal Justice Abstracts, 25, 527–546.

Eck, J.E., et al., 2005. Mapping crime: understanding hot spots. [online]. National Institute of Justice Special Report No. NCJ 209393. Washington, DC: U.S. Department of Justice. Available from: http://discovery.ucl.ac.uk/11291/1/11291.pdf [Accessed 21 May 2012].

Fabrikant, R., 1979. The distribution of criminal offenses in an urban environment: a spatial analysis of criminal spillovers and of juvenile offenders. American Journal of Economics and Sociology, 38, 31–47.

Foresman, T.W., ed., 1998. The history of geographic information systems: perspectives from the pioneers. Upper Saddle River, NJ: Prentice Hall.

Frost, G.A., 2005. Integrating GIS, GPS and MIS on the Web. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 183–196.

Getis, A. and Ord, J.K., 1992. The analysis of spatial associa- tion by use of distance statistics. Geographical Analysis, 24, 189–206.

Gorr, W.L. and McKay, S.A., 2005. Application of tracking sig- nals to detect time series pattern changes in crime mapping systems. In F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 171–182.

Grady, S.C. and Enander, H., 2009. Geographic analysis of low birthweight and infant mortality in Michigan using auto- mated zoning methodology. International Journal of Health Geographics, 8, 10.

Groff, E.R., 2007a. Simulation for theory testing and experimen- tation: an example using routine activity theory and street robbery. Journal of Quantitative Criminology, 23, 75–103.

Groff, E.R., 2007b. ‘Situating’ simulation to model human spatio-temporal interactions: an example using crime events. Transactions in GIS, 11, 507–530.

Groff, E.R., 2008. Adding the temporal and spatial aspects of rou- tine activities: a further test of routine activity theory. Security Journal, 21, 95–116.

Grubesic, T.H., Mack, E., and Murray, A.T., 2007. Geographic exclusion: spatial analysis for evaluating the implications of Megan’s Law. Social Science Computing Review, 25, 143–162.

Grubesic, T.H. and Murray, A.T., 2001. Detecting hot-spots using cluster analysis and GIS. Paper presented at the 5th annual international crime mapping research conference, 1–4 December, Dallas, TX.

Guo, D., 2008. Regionalization with dynamically constrained agglomerative clustering and partitioning (REDCAP). International Journal of Geographical Information Science, 22, 801–823.

Hall, D. and Liu, L., 2009. Cops and robbers in Cincinnati: a spatial modeling approach for examining the effects of aggressive policing. Annals of GIS, 15, 61–71.

Hammond, L. and Youngs, D., 2011. Decay functions and criminal spatial processes: geographical offender profiling

of volume crime. Journal of Investigative Psychology and Offender Profiling, 8, 90–102.

Harries, K., 1999. Mapping crime: principle and practice. Washington, DC: NIJ, US Department of Justice.

Hesseling, R.B.P., 1995. Displacement: a review of the empirical literature. In: R.V. Clarke, ed. Crime prevention studies. Vol. 2. Monsey, NY: Criminal Justice Press, 197–230.

Jefferis, E., ed., 1999. A multi-method exploration of crime hot spots: a summary of findings. Washington, DC: National Institute of Justice.

Kent, J., Leitner, M., and Curtis, A., 2006. Evaluating the usefulness of functional distance measures when calibrat- ing journey-to-crime distance decay algorithms. Computers, Environment, and Urban Systems, 30, 181–200.

Kenwitz, J.W., 1987. Cartography in France: 1660–1848. Chicago, IL: University of Chicago Press.

Land, K.C., McCall, P.L., and Cohen, L.C., 1990. Structural covariates of homicide rates: are there any invariances across time and social space? The American Journal of Sociology, 95, 922–963.

LeBeau, J.L. and Leitner, M., 2011. Introduction: progress in research on the geography of crime. The Professional Geographer, 63, 161–173.

Leipnik, M.R., et al., 2003. Apprehending murders in Spokane, Washington using GIS and GPS. In: M.R. Leipnik and D.P. Albert, eds. GIS in law enforcement: implementation issues and case studies. Boca Raton, FL: CRC Press, 168–183.

Levenson, J.S. and Cotter, L.P., 2005. The impact of sex offender residence restrictions: 1000 feet from danger or one step from absurd. International Journal of Offender Therapy and Comparative Criminology, 49, 168–178.

Levine, J.P., 1976. The potential for crime overreporting in crimi- nal victimization surveys. Criminology, 14, 307–330.

Levine, N. and Block., R., 2011. Bayesian journey-to-crime estimation: an improvement in geographic profiling method- ology. The Professional Geographer, 63, 230–243.

Liu, L. and Eck, J., eds., 2008. Artificial crime analysis systems: using computer simulations and geographic information sys- tems. Hershey, NY: IGI Global.

Liu, L., et al., 2005. Simulating crime events and crime patterns in a RA/CA model. In: F. Wang, ed. Geographic informa- tion systems and crime analysis. Hershey, PA: Idea Group Publishing, 197–213.

Lowry, S.S. and La Vigne, N., 2011. Measuring potential diffu- sion of benefits and crime displacement near public surveil- lance systems. Geography & Public Safety, 3 (1), 9–12.

Mamalian, C., LaVigne, N., and Groff, E., 1999. The use of com- puterized crime mapping by law enforcement: survey results (NIJ Research Preview). Washington, DC: National Institute of Justice.

McCarthy, T. and Ratcliffe, J.H., 2005. Garbage in garbage out. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 45–59.

McCue, C., 2011. Proactive policing: using geographic analysis to fight crime. Geography & Public Safety, 2 (4), 3–5.

Messner, S.F., et al., 1999. The spatial patterning of county homi- cide rates: an application of exploratory spatial data analysis. Journal of Quantitative Criminology, 15, 423–450.

Miethe, T.D., 1991. Citizen-based crime control activity and victimization risks: an examination of displacement and free- rider effects. Criminology, 29, 419–439.

Morenoff, J.D., Sampson, R.J., and Raudenbush, S.W., 2001. Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology, 39, 517–559.

Mu, L. and Wang., F., 2008. A scale-space clustering method: mitigating the effect of scale in the analysis of zone-based

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

170 F. Wang

data. Annals of the Association of American Geographers, 98, 85–101.

Nisbett, R.E., 1993. Violence and U.S. regional culture. American Psychologist, 48, 44–149.

North, M.J., Collier, N.T., and Vos, J.R., 2006. Experiences cre- ating three implementations of the Repast agent modeling toolkit. ACM Transactions on Modeling and Computer Simulation, 16, 1–25.

Openshaw, S., 1977. A geographical solution to scale and aggregation problems in region-building, partitioning, and spatial modelling. Transactions of the Institute of British Geographers NS, 2, 459–472.

Park, R.E., 1915. The city: suggestions for the investigation of human behavior in the city environment. American Journal of Sociology, 20, 577–612.

Pauly, G.A., McEwen, T., and Finch, S., 1967. Computer map- ping – a new technique in crime analysis. In: S.A. Yefsky, ed. Law enforcement science and technology. Vol. 1. New York: Thompson Book Company.

Radil, S.M., Flint, C., and Tita, G.E., 2010. Spatializing social networks: using social network analysis to investi- gate geographies of gang rivalry, territoriality, and violence in Los Angeles. Annals of the Association of American Geographers, 100, 307–326.

Ratcliffe, J.H., 2003. Intelligence-led policing [online]. Trends and Issues in Crime and Criminal Justice Series No. 248. Australian Institute of Criminology. Available from: http:// www.aic.gov.au [Accessed 21 May 2012].

Ratcliffe, J.H., 2004. Geocoding crime and a first estimate of a minimum acceptable hit rate. International Journal of Geographical Information Science, 18, 61–72.

Ratcliffe, J.H. and Breen, C., 2011. Crime diffusion and displace- ment: measuring the side effects of police operations. The Professional Geographer, 63, 230–243.

Rattner, A. and Portnov, B.A., 2007. Distance decay function in criminal behavior: a case of Israel. Annals of Regional Science, 41, 673–688.

Reaves, A.L., 1992. The cultural ecology of rural white homicide in the southern United States. Dissertation (PhD). University of Michigan.

Reinier, G.H., et al., 1977. Crime analysis in support of patrol – National Evaluation Program, phase 1 report. Washington, DC: US Department of Justice, National Institute of Justice. NCJ 39887.

Rengert, G.F., Piquero, A.R., and Jones, P.R., 1999. Distance decay reexamined. Criminology, 37, 427–445.

Robinson, A.H., 1982. Early thematic mapping in the his- tory of cartography. Chicago, IL: University of Chicago Press.

Rogers, E.M., 1995. Diffusion of innovations. New York: Free Press.

Roncek, D.W. and Montgomery, A., 1995. Spatial autocorrelation revisited: conceptual underpinnings and practical guidelines for the use of the generalized potential as a remedy for spatial autocorrelation in large samples. In: C.R. Block, M. Dabdoub, and S. Fregly, eds. Crime analysis through com- puter mapping. Washington, DC: Police Executive Research Forum, 99–110.

Rossomo, K., Laverty, I., and Moore, B., 2005. Geographic profil- ing for serial crime investigation. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 102–117.

Shaw, C.R., 1929. Delinquency areas. Chicago, IL: University of Chicago Press.

Shaw, C.R. and McKay, H.D., 1942. Juvenile delinquency in urban areas. Chicago, IL: University of Chicago Press.

Sherman, L., Buerger, M., and Gartin, P., 1989. Repeat call address policing: the Minneapolis RECAP experiment. Final report to the National Institute of Justice. Washington, DC: Crime Control Institute.

Smith, S.C., 2011. History, current state and future trends in GIS in law enforcement and corrections in the US. Keynote speech at the National Conference on GIS Applications in Police, 19 July 2011, Beijing.

Stigler, S.M., 1997. Adolphe Quetelet. In: N.L. Johnson and S. Kotz, eds. Encyclopedia of statistical sciences. New York: John Wiley & Sons, 64–66.

Sutherland, E., 1924. Principles of criminology. Chicago, IL: University of Chicago Press.

Swaine, J., 2008. Serious violent crime under-reported for a decade. The Telegraph, 23 Oct 2008.

Taylor, P.E. and Huxley, S.J., 1989. A break from tradition for the San-Francisco Police – patrol officer scheduling using an optimization-based decision support system. Interfaces, 19, 4–24.

Uchida, C.D., 2010. A national discussion on predictive policing: defining our terms and mapping successful implementation strategies. Washington, DC: US Department of Justice. NCJ 230404.

US Federal Bureau of Investigation (FBI), 1944. Spot maps in crime prevention. FBI Law Enforcement Bulletin. Washington, DC: US Department of Justice.

Van Koppen, P.J. and De Keijser, J.W., 1997. Desisting dis- tance decay: on the aggregation of individual crime trips. Criminology, 35, 505–515.

Wang, F., ed., 2005a. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing.

Wang, F., 2005b. Job access and homicide patterns in Chicago: an analysis at multiple geographic levels based on scale- space theory. Journal of Quantitative Criminology, 21, 195–217.

Wang, F., 2006. Quantitative methods and applications in GIS. Boca Raton, FL: CRC Press.

Wang, F., 2012. Measurement, optimization and impact of healthcare accessibility: a methodological review. Annals of the Association of American Geographers. DOI: 10.1080/00045608.2012.657146.

Wang, F. and Arnold, M., 2008. Relative deprivation, concen- trated disadvantage and homicide. Applied Geography, 28, 259–270.

Wang, F. and Minor, W.W., 2002. Where the jobs are: employment access and crime patterns in Cleveland. Annals of the Association of American Geographers, 92, 435–450.

Wang, F. and O’Brien, V., 2005. Constructing geographic areas for analysis of homicide in small populations. In: F. Wang, ed. Geographic information systems and crime analysis. Hershey, PA: Idea Group Publishing, 84–101.

Wasserman, I.M. and Stack, S., 1993. Spatial autocorrela- tion patterns with regard to suicide in the United States. In: C.R. Block and M. Dabdoub, eds. Workshop on crime analysis through computer mapping proceedings. Chicago, IL: Illinois Criminal Justice Information Authority, 69–75.

Weisburd, D., 2005. Hot spot policing experiments and crimi- nal justice research: lessons from the field. Annals of the American Academy of Political and Social Science, 599, 220–245.

Weisburd, D., 2008. Place-based policing [online]. Ideas in American Policing, 9, 1–16. Available from: www. policefoundation.org/pdf/placebasedpolicing.pdf [Accessed 21 May 2012].

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5

Annals of GIS 171

Weisburd, D. and Lum, C., 2005. The diffusion of computer- ized crime mapping in policing: linking research and practice. Police Practice and Research, 6, 419–434.

Weisburd, D. and Mazerolle, L.G., 2000. Crime and disorder in drug hot spots: implications for theory and practice in policing. Police Quarterly, 3, 331–349.

Weisburd, D. and McEwen, T., 1997. Introduction. In: D. Weisburd and T. McEwen, eds. Crime mapping and crime prevention. Monsey, NY: Criminal Justice Press, 1–23.

Wilson, R.E., 2007. The impact of software on crime mapping. Social Science Computing Review, 25, 135–142.

D ow

nl oa

de d

by [

T ex

as A

& M

U ni

ve rs

it y

C or

pu s

C hr

is ti

] at

1 5:

54 1

3 A

pr il

2 01

5