Source Of The Criminology Data

Published Date: 02 Nov 2017

Methodology

To confirm our hypothesis we will use the following methodology. First we obtain data on number and location of violations against public order. Those data are spatially referenced and projected in a map as points, and as well each public space we analyse is assigned with number of violations it bears. Because the study focuses on the relation of violations and quality of public spaces we assign each public space a numeric value representing its quality. The evaluation of public spaces has to happen beforehand.

At this point all the necessary data are applied to a map, spatially referenced and we can execute chosen analyses. Those analyses are performed using software solutions described in corresponding chapter, mainly ArcGIS and GeoDA.

Source of the Criminology Data

To analyze and predict crimes we need criminal data. Criminal data are usually obtained by various crime fighting organizations primarily national, local and metropolitan police departments, national law enforcement organizations (e.g. USA FBI) and international law enforcement organizations (e.g. Interpol, Europol).

The data come from several sources: emergency calls (911, 999, 112 and other country specific emergency lines which usually record the calls), acts reported in person, acts observed by law enforcement officers (in person or via CCTV) and surveys. Obtaining crime data is a difficult task for crime itself being secretive. Because of that only crime or violation that has been noticed or betrayed is recorded to law enforcement systems. Even when a crime has been noticed it might not get recorded. Either because the beholder decided not to disclose it or the law enforcement officer decided not to record it. Not disclosing a crime happens especially with sexual offends and child abuse because of shame, embarrassment or fear that beholders might feel. Often law enforcement officers do not report on a crime because they are being persuaded not to, or they decide pragmatically not to do it (e.g. taking into account children of offenders and other factors). Victimization surveys help to determine if crime was recorded correctly and might provide further data about individual observations. (Catalano, 2006)

After being recorded the data are kept in law enforcement systems typically computer databases. A big drawback of current systems is that they are very often not interconnected thus resulting in unanalysable data. The data becomes unanalysable because frequently jurisdictions of the various law enforcement organizations overlap and this leads to incomplete record of crime acts or violations.

The crime data are generally publicly available information (depends on national or federal laws). But in many cases a citizen has to contact a law enforcement organization and ask for specific data to obtain it. That is the case of data for this study I was requesting from metropolitan police of Bratislava. But there are some examples of data available freely on the Internet. The most of the statistics and data available are not with spatial reference or at least not compatible with common GIS solutions. I would like to bring to your attention that the research was done in English and Slovakian language so there might be more examples of available criminal data on the Internet in differently speaking countries.

International Availability

Federal Bureau of Investigation (of The United States of America) has a section in its website with many reports and statistics on U.S. national criminality. Uniform Crime Reports is an IT tool for public (http://www.ucrdatatool.gov/) led by FBI gathers crime statistics from law enforcement agencies across the U.S. Nation that voluntarily participates in the Uniform Crime Reporting (UCR) Program since 1930. Each year the data are published and are also available on their website. The data available do not have spatial reference. The FBI website itself offers various crime data e.g. Law Enforcement Officers Killed and Assaulted, Bank Crime Reports, Campus Attacks, Financial Crimes Report, Financial Institution Fraud and Failure Reports etc. (U.S. government, U.S. Department of Justice, 2012)

Another component of the U.S. Department of Justice is Bureau of Justice Statistics founded in 1979 by Justice Systems Improvement Act of 1979, Public Law 96-157 (the 1979 Amendment to the Omnibus Crime Control and Safe Streets Act of 1968, Public Law 90-351). The mission of the organization, as its website state, is to collect, analyze, publish, and disseminate information on crime, criminal offenders, victims of crime, and the operation of justice systems at all levels of government. The institution offers many statistics on courts, law enforcement, victims of crimes and various crime types on both federal and state level. The data available do not have spatial reference; mostly the data are reports rather than data ready for further analysis and are mostly in PDF or plain text format. (U.S. Department of Justice)

The National Archive of Criminal Justice Data (NACJD) of the U.S. offers much more sophisticated database (http://www.icpsr.umich.edu/icpsrweb/NACJD/). The mission of NACJD is to facilitate research in criminal justice and criminology, through the preservation, enhancement, and sharing of computerized data resources; through the production of original research based on archived data; and through specialized training workshops in quantitative analysis of crime and justice data. Data formats available are plain text (ASCII), SAS, SPSS, Stata and GIS compatible formats. Some of the not GIS compatible data have spatial reference but has to be manually referenced in a GIS. The GIS compatible formats are available only for four data collections on very specific topics and specific locations:

Development of Crime Forecasting and Mapping Systems for Use by Police in Pittsburgh, Pennsylvania and Rochester, New York 1990-2001 (ICPSR 4545)

Examination of Crime Guns and Homicide in Pittsburgh, Pennsylvania, 1987-1998 (ICPSR 2895)

Exploratory Spatial Data Approach to Identify the Context of Unemployment-Crime Linkages in Virginia 1995-2000 (ICPSR 4546)

Geographies of Urban Crime in Nashville, Tennessee, Portland, Oregon, and Tucson, Arizona 1998-2002 (ICPSR 4547)

Software company Esri, producer of the most used GIS solution ArcGIS, is offering for a payment not only software solutions but also data to its clients. Such offering is product Crime Indexes a collection of data on major categories of personal and property crime in the U.S., annually updated, not equally rich and comprehensive across the U.S.

One of especially worth mentioning projects is the Philadelphia Police Departments Philadelphia NIS CrimeBase (http://cml.upenn.edu/crimebase/), part of Philadelphia Neighborhood Information System (NIS). The website offers many tables loaded with data on various criminal acts, reports and also online maps in which user can define what will be displayed, but what is really unique is the way the database is updated daily within 5 days after an incident has happened. On the other hand not every function of the website takes advantage of it, for example maps are usually not newer than from the year 2010. Some of the raw data can be accessed via Pennsylvania Spatial Data Access (http://www.pasda.psu.edu/default.asp) a project housing geospatial data of Pennsylvania.

In most of the examples the data lacked any spatial reference besides being associated with a state, in cases when spatial reference was present it was almost exclusively in a form of a textual field. In all examples (except the Crime Indexes by Esri which I haven’t had the chance to examine) the data with GIS compatible reference available were incomplete and outdated. From a citizen’s point of view there is no option to take a look at criminality at local level (e.g. city or neighborhood scale).

Availability in Slovak Republic

Criminal data in Slovakia is recorded by national police, military police, railway police, justice and prison guard, customs police and various metropolitan law enforcement organizations. There is no central system to gather criminality data. The data are usually space referenced only in textual form not in a form of coordinates or similar.

Most comprehensive source on criminality data in the country is Statistical Office of the Slovak Republic which includes all the above mentioned sources except local / metropolitan law enforcement organizations. The office provides for a payment reports on criminality on national or regional level. There are free data and statistics available at their website (http://portal.statistics.sk/showdoc.do?docid=48) available either as publications or so called predefined tables. The statistics and data include all the crimes recorded by a law enforcement organization according to act. no 300/2005 of criminal law. Criminality according to the act is divided into these three types:

general criminality (which further divides):

property criminality

violence

moral criminality

other general criminality

economical criminality

other criminality

Under section publications we can find two options print and electronic publications. Print publications are available for a payment by person in the office or through the post. They cover period of years 2005 to 2007. Where print publications end the electronic ones begin covering years 2008 – 2009. The electronic publications are available for free and describe criminality on national and regional level. The reports are very general and appropriate mostly for analysis on national level.

Predefined tables include very general data about distribution of crime among various types of crime and their evolution over time. The data is ranging from 2003 to 2011.

Another source on crime data available online is Ministry of Interior of the Slovak Republic and their website. The ministry, besides other competences, is in charge of national police, justice and prison guard, and intelligence agencies and these are probable sources for data available on their official website. Statistics and data available are for years 1997 – 2012, while I was doing research for this study in January 2013 there were already available statistics for the whole year 2012. The data are organized by year, and are divided by type of crime, name of the criminal act and the paragraph of an act the criminal act belongs to. Also each of the years is further sorted into general crimes, criminal acts on youth, criminal acts committed by foreigners, and at the same time sorted by regions.

Finally there is General prosecution of the SR which includes on its website statistics on cases, criminal acts and other violations of law from perspective of prosecution. Each year the office publishes a year report containing all the mentioned statistics on over 200 pages (http://www.genpro.gov.sk/statistiky-12c1.html) since 1999 and the newest available report is from year 2011.

Criminology Data for the Purposes of This Study

The data used in this research for calculations and analyses were provided by the metropolitan police of the capital of Slovak republic – Bratislava.

The metropolitan police is entitled to issue fines in strictly given cases by act no. 372/1990 about violations (e.g. against entrepreneurs, against traffic safety and fluency, against culture etc.) . This study is focusing solely on violations of § 47 dealing with violations against public order.

Offences against public order

§ 47

(1) A person is guilty of an offence if he-

a) does not obey an order of a constable on duty

b) generates noise at night,

c) generates public indecency,

d) pollutes public areas, sites open to public or defiles objects with fliers, advertisement or neglects his duty of cleaning of public areas,

e) intentionally destroys, damages, defiles or unlawfully removes, alters, modifies, moves or covers the tourist guide or any other public sign,

f) violates rules for protection of public order during public sport or cultural events or such rules in places for recreation or tourism,

g) damages or unlawfully occupies a public area, a publicly accessible building or a public facility

h) wears in publicly accessible areas cold weapons, especially knives, bayonets and cutlasses, unless they are a part of a uniform, national or historical costumes, equipment and armaments of armed forces and armed security forces, or a sport activity, lawful hunting, lawful fishing, profession or employment, as well as other objects that may harm the health, if it is possible from the circumstances of the case or conduct of the person to deduce that they are to be used for violence or threat of violence.

ch) uses fireworks contrary to the general regulations or instructions of the fireworks.

(2) For an offense under section 1 subsection a) to d) there is a fine of up to 1,000 SK, for an offense under subsections 1. e) to g), a fine of 3,000 SK and for an offense under subsections 1. h) a fine of up to 5 000 SK. [1] 

Introduction to the Study Area of Lamač

Lamač is a part of, capital of Slovak republic, city of Bratislava. By area size of 654ha it is the smallest city part (city tract) with population of 6745. (Štatistický úrad Slovenskej republiky)

Lamač is located at north of Bratislava only less than 7km from the city centre. In touch with freeway D2 connecting Hungary and Czech Republic easing access to other parts of Bratislava. This part of city also features a small railway station with direct connection to Austria and city centre where further rail connections are available. Local public mass transport system operates in Lamač with five bus lines using 15 bus stops with relatively big gaps in intervals. On north and east Lamač neighbours with forests and there is in general very good access to nature.

History of Lamač

The oldest evidence of settlement in the area of Lamač dates back to 13th century. Lamač itself was founded in 1506 supposedly by Ján Skerlič. The village was firstly known as Krabatendorff (Croatian village) because of population consisting mostly of Croatian colonists; later it became known as Blumenau taking its name from a former village of that name in north. In 1549 the name Lamas comes to use. Year 1624 brought Turkish army which plundered the village.

Lamač was in regression in next years and couldn’t afford neither a vicar nor a teacher. In years 1643 to 1752 it was annexed to near village of Záhorská Bystrica. A plague came to the region in 1679. From 1703 to 1711 Lamač was four times plundered. Population was in decline, and another plague in 1714 only facilitated the decline.

Since 1752, when the village became independent, the population and the village itself started to be on the rise. A fire and another plague affected the village in 18th century. In 1846 had begun construction of a railway bringing also disease of typhus. On 22nd of July 1866 a decisive battle of the Austro-Prussian War took place near Lamač. The post office opened for the first time here in 1922.

During the World War II Lamač was supposed to be annexed to the Third Reich. Lamač was liberated by the Red Army on 5th of April 1945. In 1946 on 1st of April was the village annexed to Bratislava becoming its part. During the 1970s a big construction of soviet housing development raised population of Lamač to the numbers similar to todays; this part of Lamač is called "New Lamač" and houses the most of the neighborhood’s services. (Mestská časť Bratislava - Lamač)

Demographics and Census Data

Criminality of Lamač

Lamač is seen in public eyes as a quiet and calm neighbourhood very popular among young families looking for a place to raise their children especially to raise them in Christian faith considering the number of believers in the area. Since 2006 is Lamač equipped with two CCTV cameras at Malokarpatské square. (Hlavné mesto SR Bratislava, 2007)

On the other hand Lamač was for a very long time home to many important bosses of national organized crime such as "Piťovci" mafia clan. On 15th of November 2011 at 4a.m. the up to now biggest police action called "Eden" took place in Lamač. A helicopter and many vehicles of SWAT team were included in the action to arrest the boss of the clan Juraj Ondrejčák. (meg, 2011)

In 1997 a house in Lamač happened to be a crime scene for a brutal murder and torment committed by four homeless people to two other homeless men. For one week the victims were tied up, didn’t get any food, got almost only pure alcohol to drink, were drugged, burned several times, beaten by the offenders many times with an iron bar and a brick and finally castrated. One victim died of hunger the other one was strangled. (vr, 2012)

This study is focusing solely on violations of metropolitan ordinance of § 47 of act no. 372/1990 of 28 August 1990 dealing with violations of public order.

Methods of Evaluation of Public Spaces

Possible Criteria for Evaluation of Public Spaces

In order to confirm the set hypothesis first we need to assign a numerical value to each open space in the study area. To obtain this numerical value we must evaluate each public space on several level (by several criteria).

Stiles (2011) provides a set of criteria for assessment of open public spaces. These are divided into three categories based on their functions. The categories and their corresponding functions are in a table below.

Environmental and ecological functions

Climatic amelioration

Noise screening

Influencing the hydrological cycle – storm water management

Providing habitats for wild plants and animals

Social and societal functions

Providing space and facilities for leisure and recreation

Facilitating social contact and communication, including cultural and commercial activities

Allowing access to and experience of nature

Influencing human physical and psychological health and well-being

Structural and symbolic functions

Articulating, dividing and linking areas of the urban fabric

Improving the legibility of the city or neighbourhood

Establishing a sense of place

Acting as a carrier of identity, meanings and values

The criteria are meant for assessment of future projects and are not designed for the purposes of this study. The hypothesis of this paper is partly built upon the broken windows theory and studies the applicability of the theory on study area and its part which suggests that disorder leads to more disorder before it leads to crime. The purpose of this study is not to prove whether disorder leads to crime; rather whether disorder in form of poor urban design and state of the spaces leads to more disorder. For the most objective and relevant evaluation to our study subject, we enrich the mentioned criteria of few more: presence of waste control features, long-term state of present greenery, presence of CCTV, quality of security features of urban design. Some original criteria are omitted as not important in our study and based on the changes new set criteria was compiled. The resulting criteria used in this study are in table…

Environmental and ecological criteria

Noise screening (used materials, height of greenery and ratio of evergreen plants)

Providing habitats for wild plants and animals (linkage to other green areas, occurrence of miscellaneous plants and animals)

Presence of waste control features (number of the features such as litter bins and dog litter bins)

Long-term state of present greenery (quality of greenery throughout the whole duration of the study)

Social and societal criteria

Providing space and facilities for leisure and recreation (number and presence of the facilities)

Facilitating social contact and communication, including cultural and commercial activities (amount of people spontaneously communicating)

Allowing access to and experience of nature (ease of access to natural features)

presence of CCTV (number of present CCTV units)

quality of security features of urban design (mostly visibility at the public space)

Structural and symbolic criteria

Articulating, dividing and linking areas of the urban fabric (blend of various functions)

Improving the legibility of the city or neighbourhood (ease of movement in the space, orientation clarity)

Acting as a carrier of identity, meanings and values (level of everyday use)

Weighting of the Criteria for Evaluation of Public Spaces

Because not every criterion is equally important the spaces are not assigned points based on simple arithmetic mean. Each criterion has to be assessed first of how important it is for the evaluation. For this purpose we utilise an analytic hierarchy process method of decision making developed by Saaty (1977).

The method consists of two steps. The first step is to compare each individual criterion with each other. Every couple is assigned with a number based on which one and how much is more important than the other. (see table …) Then those values become input of a matrix which calculates weights suitable for weighted mean.

The importance values are assigned based on this table:

Intensity of Importance

Definition

Explanation

1

Equal Importance

Two activities contribute equally to the objective

2

Weak or slight

3

Moderate importance

Experience and judgement slightly favour one activity over another

4

Moderate plus

5

Strong importance

Experience and judgement strongly favour one activity over another

6

Strong plus

7

Very strong or demonstrated importance

An activity is favoured very strongly over another; its dominance demonstrated in practice

8

Very, very strong

9

Extreme importance

The evidence favouring one activity over another is of the highest possible order of affirmation

Reciprocals of above

If activity i has one of the above non-zero numbers assigned to it when compared with activity j, then j has the reciprocal value when compared with i

A reasonable assumption

1.1–1.9

If the activities are very close

May be difficult to assign the best value but when compared with other contrasting activities the size of the small numbers would not be too noticeable, yet they can still indicate the relative importance of the activities.

(Decision making with the analytic hierarchy process, 2008)

Each pair of criteria is assigned a value and calculated using the Saaty matrix. The resulting weights used for our mean are as follows:

 

A

B

C

D

E

F

G

H

I

J

K

L

 

 

 

A

1

7

1/7

1/7

1/7

1/7

4

1/8

1/7

1/8

1/2

1/8

1,62693E-06

0,329316878

0,018748624

2%

B

1/7

1

1/9

1/7

1/9

1/9

1/6

1/9

1/9

1/9

1/8

1/9

8,88925E-11

0,145346784

0,008274863

1%

C

7

9

1

2

4

4

7

1/3

1/3

4

6

3

112896

2,636674665

0,150110804

15%

D

7

7

1/2

1

3

3

7

1/2

1/4

3

6

3

10418,625

2,161810062

0,123075877

12%

E

7

9

1/4

1/3

1

1/7

2

1/3

1/4

1

4

1

0,5

0,943874313

0,053736524

5%

F

7

9

1/4

1/3

7

1

8

2

2

3

7

1

24696

2,323016091

0,132253637

13%

G

1/4

6

1/7

1/7

1/2

1/8

1

1/9

1/9

1/8

1/7

1/8

5,27245E-08

0,247457724

0,01408823

1%

H

8

9

3

2

3

1/2

9

1

1/2

1/2

4

1/2

2916

1,944161297

0,110684727

11%

I

7

9

3

4

4

1/2

9

2

1

4

6

3

1959552

3,34461705

0,190415284

19%

J

8

9

1/4

1/3

1

1/3

8

2

1/4

1

6

1/2

24

1,303219601

0,074194721

7%

K

2

8

1/6

1/6

1/4

1/7

7

1/4

1/6

1/6

1

1/8

9,64506E-05

0,462763131

0,026345968

3%

L

8

9

1/3

1/3

1

1

8

2

1/3

2

8

1

682,6666667

1,722598474

0,098070742

10%

 

 

 

 

 

 

 

 

 

 

 

 

 

SUM:

17,56485607

1

The next step in the process of evaluation of the public spaces is to assign points to each space based on given criteria. The result is a set of points for each space. Using weighted mean with calculated weights we are able to evaluate each space with a single point value. Those values are after that placed on map corresponding to their actual location.

Suitable Methods of Spatial Analysis

Methods used to reveal relations in the placement of public order violations and the quality or state of the public space in the same location include following... Spatial autocorrelation is used to determine dependency of placement of individual observations of a certain class of the violations to placement of other observations of the same class of violations. Spatial regression is used in the study to determine dependency of the individual observations of a certain class of the violations to observations of other class of the violations. Finally LISA and kernel smoothing is used to graphically represent the placement and relations among the studied observations and to simplify perception of the studied data.

Global Indicators of Spatial Association

Indicators of spatial association in general are statistics that evaluate the existence of clusters in the placement of observations within one file. Existence of clusters is a sign of not random spatial arrangement.

A fundamental concept in the analysis of spatial autocorrelation for areal data is the spatial weights matrix. This is a square matrix of dimension equal to the number of observations, with each row and column corresponding to an observation. (Anselin, et al., 2000)

Spatial autocorrelation statistics measure and analyse the degree of dependency between pairs of locations by comparing their spatial weights. Classic test statistics for spatial autocorrelation are the join count statistic, Moran’s I and Geary’s c. (Cliff, et al., 1973) Spatial autocorrelation can be either more positive than expected from random indicating the clustering of similar values, or significantly negative spatial autocorrelation indicates that neighbouring values are more dissimilar than in random distribution. (Wikipedia contributors, 2013)

Model of spatial autocorrelation:

Xi = , i = 1, 2, ..., n,

i i = 1, 2, ..., n – independent random variables with common scatter of 2

wij, i, j = 1, 2, ..., n – known constants, bound to locations i and j

 – constant of degree of spatial autocorrelation

In this study I utilise global Moran’s I (via GeoDA and ArcMap) to determine the degree of autocorrelation:

I = ,

n – number of areas

A – number of boundaries

ij = 1 – areas are neighbours

xi (i = 1, 2, ..., n) – values of the observation in the area i

I → +1 – positive autocorrelation

I → -1 – negative autocorrelation

This approach of analysis yields only one statistic to determine whether the observations in the study area have autocorrelation or not. Therefor the analysis is called global spatial autocorrelation because there is only one number for the whole study area.

Local Indicators of Spatial Association

The dependency and autocorrelation of observations usually vary greatly across the study areas. To determine whether there is a local autocorrelation not reflected in the global autocorrelation I use local indicators of spatial association so called LISA. LISA graphically highlights the most significant areas for the spatial autocorrelation.

Local versions of Moran’s I and Geary’s C first used by Anselin (1995) can be utilised to execute the calculations.

I use Local Moran’s I incorporated in GeoDA:

Ii (d) = ,

where d is chosen critical distance.

Spatial Regression

Methods of spatial regression estimate level of spatial dependency using regression analysis and this way they provide information on spatial relationships among the variables. Spatial dependency can be result of relationship of variables of which one is independent and the other one is dependant, between the dependent variables and a spatial lag of itself, or in the error terms.

These analyses for the purpose of this study were executed in GeoDA

Kernel Smoothing

In statistics, kernel smoothing (kernel density estimation) is a non-parametric way to estimate the probability density function of a random variable. In spatial analysis kernel smoothing helps graphically represent the spatial arrangement of individual observations while highlighting the most significant areas and suppresses the least significant areas.

Kernel density estimator is:

where K(…) is the kernel.

In this study Epanechnikov kernel is utilised, this type of kernel is incorporated into ArcMap software which I use for this particular analysis:

Available Software Solutions

Spatial econometric methods are not routinely incorporated in commercial software packages. But spatial analysis is starting to be supported by many software developers. Specialized software solutions for spatial crime analysis are typically distributed solely inside the law enforcement organizations that also own it or even develop it themselves. In this section we will take a look at some of the available solutions to public and spatial / urban planners.

Besides others among software programs developed for spatial analysis are MINITAB developed by Griffith, SpaceStat developed by Anselin, which is especially suitable for spatial economy. Available are also macros for SPSS. Some of spatial models can be executed in Fortran. SAS (Statistical Analysis System) contains some procedures of spatial statistics and user has the ability to install macros specialized to methods of spatial analysis.

R

R is an open source programming language and software environment for statistical computing and graphics. R is an implementation of the S programming language combined with lexical scoping semantics inspired by Scheme. S was created by John Chambers while at Bell Labs. R was created by Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, and now, R is developed by the R Development Core Team, of which Chambers is a member. R is named partly after the first names of the first two R authors (Robert Gentleman and Ross Ihaka), and partly as a play on the name of S. (Hornik, 2012)

R provides a wide variety of statistical and graphical techniques, including linear and nonlinear modeling, classical statistical tests, time-series analysis, classification, clustering, and others.

ArcGIS

Most prevalent geographic information system is ArcGIS by Esri. It has wide usage including creation and usage of maps, compilation of geographic data, analysis of mapped information, sharing and discovering of spatial information, and management of geographic databases. ArcGIS for Desktop consists of several integrated applications, including ArcMap, ArcCatalog, ArcToolbox, and ArcGlobe.

ArcToolbox is a component which includes spatial analysis tools such as Spatial Analyst which can:

Create, query, map, and analyze cell-based raster data.

Perform integrated raster/vector analysis.

Derive new information from existing data.

Query information across multiple data layers.

Fully integrate cell-based raster data with traditional vector data sources.

Or it also includes Spatial statistics which allows you to:

Summarize the key characteristics of a distribution.

Identify statistically significant spatial clusters (hotspots/cold spots) and spatial outliers.

Assess overall patterns of clustering or dispersion.

Model spatial relationships.

These tools by themselves might be sufficient for a basic spatial crime analysis.

GeoDa

GeoDa is a solution focused particularly on spatial data analyses but also provides geovisualization, spatial autocorrelation and spatial modeling. The package was initially developed by the Spatial Analysis Laboratory of the University of Illinois at Urbana-Champaign under the direction of Luc Anselin. Development continues at the GeoDa Center for Geospatial Analysis and Computation at Arizona State University. (Anselin, 2005)

GeoDa works with shapefile format developed by Esri. GeoDa can produce histograms, box plots, and scatterplots to conduct simple exploratory analyses of the data. A tool present in GeoDa, Moran scatterplot, depicts a standardized variable in the x-axis versus the spatial lag of that standardized variable. The spatial lag is nothing but a summary of the effects of the neighboring spatial units. That summary is obtained by means of a spatial weights matrix, which can take various forms, but a very commonly used is the contiguity matrix. The contiguity matrix is an array that has a value of one in the position (i, j) whenever the spatial unit j is contiguous to the unit i. For convenience that matrix is standardized in such a way that the rows sum to one by dividing each value by the row sum of the original matrix.

LISA analysis allows us to identify where are the areas high values of a variable that are surrounded by high values on the neighboring areas i.e. what is called the high-high clusters. Concomitantly, the low-low clusters are also identified from this analysis.

Is a freestanding solution and does not require any GIS program for its functionality, on the other hand the only input can be Esri shapefile. The solutions main functions can be classified into six categories:

Spatial data manipulation and utilities: data input, output, and conversion.

Data transformation: variable transformation and creation of new variables.

Mapping: choropleth maps, cartogram, and map animation.

Exploratory Data Analysis (EDA): statistical graphics.

Spatial autocorrelation: global and local spatial autocorrelation with inference and visualization.

Spatial regression: diagnostics and maximum likelihood estimation of linear spatial regression models.

GeoDa offers different views on the same data in different windows containing maps, graphs and tables. (so called dynamic linking and brushing). Highlighting one observation in one view higlights the same observation in another window be it map, scatter plot, box plot, table etc.

The full menu bar contains 12 items. Four are standard Windows menus: File(open and close files), View(select which toolbars to show), Window(select or rearrange windows) and Help(not yet implemented). Specific to GeoDa are Edit (manipulate map windows and layers), Tools(spatial data manipulation), Table (data table manipulation), Map(choropleth mapping and map smoothing), Explore (statistical graphics), Space(spatial autocorrelation analysis), Regress(spatial regression), and Options (application-specific options).

In field of exploratory data analysis the program offers box plots, box plot maps and percentile maps to describe the overall distribution of observations. Box plot is a tool to graphically depict groups of numerical data through their value distribution: first quartile (Q1, lowest 25 percent of all observations), second quartile (Q2, next 25 percent), median, third quartile (Q3), fourth quartile (Q4), and upper outliers. Box plot map depicts the same data as box plot but located in a map differentiated by color. Another EDA method is percentile map.

GeoDa provides also tools for constructing spatial weights and tests for the presence of global and local spatial autocorrelation. The pattern of dependence or clustering is summarized in a single indicator such as Moran's I. Positive Moran's I signals that locations in spatial proximity are more similar than what is expected in randomness, negative indicator signals dissimilar location and app. zero indicator indicates random location of observations.

The program is useful for both exploratory and confirmatory spatial data analyses with either point or areal data, a big advantage of GeoDa compared to other solutions is dynamic linking and brushing and the fact that output can be in a map form without any need for a GIS program. The most prevalent weaknesses of the software are compatibility solely with shapefiles and narrow variety of methods for analyses of point data.

S+SpatialStats

S-PLUS is a commercial implementation of the S programming language (language from which R language originates); and features object-oriented programming capabilities and advanced analytical algorithms. S+SpatialStats is an extension for spatial analysis and was created specifically for the exploration and modeling of spatially correlated data.

CrimeStat

CrimeStat is a freely available software solution developed by Ned Levine & Associates, with funding by the U.S. National Institute of Justice (NIJ), an agency of the United States Department of Justice. CrimeStat is a Windows-based program that conducts spatial and statistical crime analysis and is designed to interface with a GIS. The program can analyze the distribution of the objects, identify hotspots, indicate spatial autocorrelation, monitor the interaction of events in space and time, and model travel behavior. (Ned Levine & Associates, 2006)

CrimeStat can input data both attribute and GIS files but requires that all datasets have geographical coordinates assigned for the objects. The basic file format is dBase (dbf) but shapefile (shp), and ASCII text files can also be read. The program requires a Primary File but many routines also use a Secondary File. CrimeStat uses three coordinate systems: spherical (longitude, latitude), projected and directional (angles). (Wikipedia contributors, 2012)

Distance can be measured as direct, indirect (Manhattan) or on a network (which also allows travel time or speed to be used). Distance units are decimal degrees for spherical coordinates and feet, meters, miles, kilometers, or nautical miles for projected coordinates. The program can create reference grids. Several routines also use the area of the geographical region for their calculations.

Spatial statistics package provided with CrimeStat is divided into four categories:

Spatial distribution or centrographic statistics: mean center, center of minimum distance, standard deviational ellipse, and Moran’s I spatial autocorrelation index.

Distance statistics: nearest neighbor analysis and Ripley’s K statistic.

Hotspot analysis routines: hierarchical nearest neighbor clustering, K-means clustering, and local Moran statistics.

Interpolation statistics: kernel density estimation routines.

Centrographics are first step in understanding distribution of observations or in our case crimes in studied area. Mean center is a method to identify central location of all incidents. Mean center is a base for the standard deviational ellipse which identifies dispersion using standard deviation of the distance and direction (orientation) of each incident location. The output produced by CrimeStat includes tabular summaries containing a number of descriptive statistics and graphical objects, which can be saved and imported as a shapefile.

Category of distance statistics includes nearest neighbor analysis and Ripley’s K statistic. Nearest neighbor statistics' output is nearest neighbor index which represents comparison of distance of each incident and its nearest neighbor with supposed distance if distribution of points was random. If the index equals 1.0 the average distance of incidents locations and random locations is the same. if the value of the index is below 1.0 the points are being clustered. Ripley’s K function similar method to the previous but compares the number of incidents within distance to a random point with expected number of incidents under random distribution. CrimeStat calculates 100 distance intervals (called bins) around each point location, counts the number of crime incidents within each interval, and compares this to the expected number for a spatially random distribution. If the mean average of number of incidents within a given distance is greater than expected under randomness the incidents are in clusters.

Routines for hot-spot analysis include hierarchical spatial clustering, K-means clustering, and local Moran statistics. Hierarchical clustering needs user to set minimum number of incidents that will be clustered together based on nearest neighbor analysis. These clusters are then grouped into bigger clusters until there is no data left to make cluster of; the grouping is also based on nearest neighbor analysis. K-means clustering is a tool which groups incidents into K number of clusters, where K is set by user. The default number of clusters assigned by CrimeStat is five. A local Moran procedure approach is based on the concept of local indicators of spatial association (LISA), in which each observation is assigned a score based on the extent to which significant clustering of similar values around that observation exists.

CrimeStat is useful for doing exploratory spatial data analyses with incident locations, and provides a number of statistical routines that vary from descriptive centrographic applications to more sophisticated nearest neighbor and spatial autocorrelation functions. On the other hand CrimeStat does not provide any functionality of modeling correlates or determinants of crime.