A SAS/ AF application for local police forces

Size: px

Start display at page:

Download "A SAS/ AF application for local police forces"

Alban Webb
6 years ago
Views:

1 A SAS/ AF application for local police forces Martine Beullens, University Center of Statistics- K. U. Lellven-BELGIUM Abstract One of the aims of each local police force is the prevention and the criminal investigation of local crime events. "Police Statistics" are a basic-work-instrument in improving the efficiency of daily streetwork of police officers in their battle against local problems of criminality. With this purpose in mind, was -at the Universuty Center of Statistics (K. U. Leuven) - a user-friendly SASI AF application developed for registration of and basic crimeanalysis. This was done in cooperation with Prof. Van Kerckvoorde (Faculty of Law - criminology - K.U. Leuven), who provided an advanced study of a statistical police registration system. The PC-program was the result of a study-project done by the Ministery of Internal Affairs., (: /:.' t;.' (. M [ r ii :, : :'. ':..,,,!.'.,.?, f :. g: [, ', : :. c-;, 1. : : ; i-,, :,,: ' ' The data-entry application allows the policeman to registrate the basic information of crime events. It is a program providing different steps in registration, starting with the simple street report. SASI AF and SAS/FSp@, with SCL made it possible to build in range control, field validation and selection lists to avoid error typing. The second part of the application is a menu-driven tool for simple statistical analysis of the registrated crime events. To get a total view of the criminality in the region, the policeman can produice one-way frequency and two-way crosstabulation tables, bar- and pie-charts and list reports. Goals Why developing a statistical instrument for the local police-forces? In Belgium, no exact and recent data on crime statistics are available. Yet, an unconditional need on adequate information about criminality exists in this country. Police statistics are basic to obtain a correct determination of structural problems and their causes. Data can be used before taking decisions in the actions policemen have to take in dealing with local crime problems and in controlling the effectivity and efficiency of the measures taken. In fact, police statistics can be considered as one of the most important work-instruments in the war against crime. Armed with current information, chiefs of local police forces can develop a local crimepolicy and proove and defend the measures they took. Once they obtained these recources, they can use them in a more efficient and effective manner. In other words, the ability to create local police-statistics is an indispensable tool in ruling crime problems. In Belgium, the police organisation provides that different police forces operate in the same region. Theirfore, it is an undispensable condition that the data- collected by each police station in the same region can be exchanged. To obtain that condition, a uniform harmonised registration program is necessary (?). \ \,, : ' t.-:. 297

2 Ex. If the local police chief wants to have a view on all the 'hold-ups' committed during a certain period in his region, the definition of 'hold-up' should be fixe and stable. Each police officer has to use the same pre scripted definition of 'hold-up'. In an easier way, police officers have to be sure that they talk about the same criminal incident if they talk about a hold-up.., To obtain a uniform unique registration-system for the three different police forces in Belgium, the Ministry of Internal Affairs paid a two year during universitary research to develop a new data-entry application for the registration of the crime incidents. The basic argumentation behind this research was the lack in Belgium of valuable and reliable statistics, because of the no existance of a uniform registration. Different tools of the instrument were discripted. The instrument had to give the policeman the ability to registrate with precision when and where a crime took place and what the juridical characteristics of the incident were. Basic data on the age, the capacities and the sex of its victims and offenders. In a second stadium, each local police-force must be capable to produce local crime-statistics in his region and compare and exchange this information with neighbour police-forces working in the same region. System Design The sytem consists of a SAS/ AF user interface, written in Screen Control Language (SCL) that serves as a front-end to SAS programs. It is developed under PC/DOS with the release 6.04 of the SAS software. In fact, the application consists of two parts or two sub-applications which can be invoked from the principal menu. The first sub-application performs the data input and validation, so this is the data-entry application. Each individual crime incident must be entered into the system by the police officer This input takes place through five fsedit sessions that are accessed once at the time. The data-entry screens look like registration forms, and bells and whistles provide useful error trapping for invalid data. The second application allows the policeman to produce some local statistical output of the collected crime data. The output can be presented as a table, a graph or a report. They can choose between a one-way frequency or two-way crosstabulation with a Chi-square test. Possible graphics are bar- and pie charts. This user interface consists of a menu system that accesses program input screens. The program input screens help the user select valid input parameters by providing selection lists, menu's and help screens. Associated with each program input screen are SAS programs containing the procedures that actually perform the given task. 298

3 Primary Menu I I Registration of crime incidents Output Setul= add new fact edit existing fact I Select data Tables Graphics Reports Groups Programs One-way frequency tables Two-way crosstabulations Summary statistics Class variable and statistics Bar charts I L Pie charts I Vertical bar charts Stacked bar charts Grouped bar charts SAS/ AF was choosen for the development because SAS/ AF has many features that make the system attractive for an end-user. Point and click acces, user friendliness, full screen capability are just a few features why the system can serve a local police force. Conversion Because some police-forces, especially the bigger ones, need strong PC's to work comfortably, the application was transported from the release 6.04 on PC/DOS to the release 6.08 under Windows. In this part, a brief description on how this conversion was done will follow. 1) SAS/AF and SAS/FSP catalogs export: create a transport file from all the entry's in the AF-catalog PROC CPORT catalog =.. file =.. i import: read this file to create the catalog under Windows with the PROC CIMPORT infile=.. catalog =.. The keys- and macro-entry's of the AF catalog could not be transported with the CPORT procedure. So, we remade the keys-entry's again in the converted catalog under Windows. To create the macro-entry's, we rerunned the macro program under Windows and made a new 299

4 catalog containing the macro entry's. 2) Formatcatalog To transfer a format-catalog from relaease 6.04 to release 6.08, the CPORT procedure can not be used. That's why we did it in two steps: - create a control output dataset under DOS: PROC FORMAT cntlout = - read this SAS-dataset with v603 engine under Windows - use this dataset as a control input dataset to create the format-catalog under Windows: PROC FORMAT cntlin =.. ; Once all the catalogs were transported, the application worked, but there were some problems: -It was not possible to make a selection by clicking with the mouse button on the appropriate choice. Therefore, we had to assign choice group names to the various screen fields in the ATTR window for each PROGRAM entry of the AF-catalog. This meant we had to re-compile the SCL-source code. -We also had to change the window sizes in the GATTR window for each PROGRAM, HELP and MENU entry of the catalog. In those windows where it was usefull to have scroll bars, we used the scroll bar option in the GATTR window. -In the exented table program we had to change the "refresh" statement with the "call execcmdo" statement containing the keyfield-command. Future Because the application is converted from release 6.04 on DOS, it does not use all the new features of the Version 6 SAS/ AF software yet. We are now busy with re-writing the SCL converted code using the new SCL programming statements. Combining the new features of choice groups with the features of extended tables makes the SCL code more simply and the application easier to maintain. A great permoformance improvement can be realized. Also the new techniques of screen design are more user-friendly and really windows-like. The appearance of the screens look more professional. Presentation Different tasks can be selected from the Primary Menu (Figure 1). Using the mouse,one has to point out the chosen field and click on it, with the keyboard one has to place the cursor on the selected field and press the ENTER key. The left button calls the data-entry application. On the right side, we have the button to go to the menu for producing statistical output. 300

Figure 1 : Primary menu The option in the middle must be used before starting with the statistics. It runs a program that creates some additional variables from the registrated variables.

5 Figure 1 : Primary menu The option in the middle must be used before starting with the statistics. It runs a program that creates some additional variables from the registrated variables. These can then be analyzed immediately in the statistical menu. l)data-entry application Let's take the example of a man who attacks his wife in their appartment and treatens her with a gun. To enter this criminal act, the policeman chooses "registration of a new incident" from the registration-menu. First of all, the exact number of the crime-report must be filled in. This is a unique administration-number identifying the file 'crime incident'. 301

6 The fsedit session consists of five separate fsedit screens that are accessed one at a time. The different fsedit sessions that were invoked instead of one big session with over 1500 lines of SCL do not only helped to solve memory problems (with 6.04 under,dos) but also made the system work quicker. After clicking OK, a serie of FSEDIT screens will be presented grouped in different modules. The screens look like registration forms on which crime information must be filled in. The incident-number appears automatically on every screen. Field verification is build in. If the allowed entries are Y (Yes) or N(no) and one types a P for example, an error message appears on the instruction part above. As one can see, the top lines of each screen are reserved for giving instructions and error messages to the user. The correct code must be entered first before continuing. When a screen is completed, with the function key F11, one moves to the next screen asking for new information to fill in. There are two possibilities to fill in the data field. If one knows the code-table, one can just type the specific code. The description appears on the screen as a kind of verification. If one does not know the allowed codes, a facility is developed for the user to type a "?" in the field to access a list of possible codes from which he can make the appropriate selection. The most important module is the second module(second data-set) which is the offence module. This screen handels about the basic information of the crime incident, which includes the type of offences that are involved in the crime (is it a robbery, an assault offence,... ). There are 3 classification-systems to registrate the offence in an exact way. The most important is the juridical classification (penal law-code). A juridical classifications in different catagories was made, following the exacts prescriptions of the penal law. For the firts time in our country the penal law code (the old book) is computerized. These global categories are presented as a list on the screen from which the user can select one item. If we return to our example of the attack of a man towards his wife, we must select the first categorie (offences against the fysical welfare, attacks and assaults). A second screen (subdivisions) appears containing all the offences belonging to that global categorie (Figure 3). The way that this penal law-code was developed can be called the 'tree structure'. This structure guides the police officer towards. the exact juridical qualification of the offence. This collection of offences (nomenclature) is stored in a SAS dataset and the records belonging to the selected categorie are picked up from that database and displayed on the screen in which the user can scroll and select again the right item. This system could be realized by the use of extended tables. An extended table is a dynamic data structure in the sense that there is no limit set to how many rows can be displayed. This was very usefull here because the number of offences belonging to a categorie depends on the categorie itself. 302

7 3 : registration of the type of offence- tree structure To indicate the location of the incident, the name of the street must be entered first. Therefore, the user has to type the first part of the streetname, and a list of all the streets appears on the screen with the first occurence of that street containing the string as the top row of the list. When we return, the community is automaticaly filled in by the screen control. On the next screens of the offence module, some other informations must be filled in, like the location type, the incident date, the time that the incident occured and whether there was any threatening or not. The third fsedit. session opens the 'object-module' which must only be filled in if the. incident was a robbery, vandalism or drug offence. For it is only is this categories of offences that the registration of the object that was stolen -or destroyed is important. The last two modules are used to enter the characteristics of the victims and suspects that were involved in the crime incident. Capacities as age, sex and nationality are registrated. If the crime was committed by more than one person. the suspect module must be repeated for each offender. To add a new suspect, a function key is provided to get a new empty screen to fill in the information of that suspect. 2)StatisticaJ application Before starting with producing tables and graphs, we have to. select the crime data we want to analyze. This must be done in the Selection window of the menu for statistical output (Figure4). At this window, the period, the type of offence, the community and the module must be specified by the user. For example, the criminal incidents occured in 1990 in the COnmiunity 303

"Leuven" obtain our special interest. We can analyze all crimes or select some of them from a selection list on the screen: f.ex. the assault offences, robbery offences, and vandalism crimes.

8 "Leuven" obtain our special interest. We can analyze all crimes or select some of them from a selection list on the screen: f.ex. the assault offences, robbery offences, and vandalism crimes. After clicking RUN, all the robbery's, assault incidents and vandalism crimes are picked up from the database and stored in a work dataset on which we are going to produce our tables and graphical displays. A report of our selections appears automatically on the output screen (Figure 5). Figure5 : Output window - report of selections 304

Once the appropriate data are selected, tables and graphs can be produced. Figure 6 is the program input screen to produice a two-way frequency table.

9 Once the appropriate data are selected, tables and graphs can be produced. Figure 6 is the program input screen to produice a two-way frequency table. After selecting the row and column-variable, the report can be generated immediately by clicking on the RUN button. Figure 6 : program input screen for a two-way frequency table A lot of additional options are available but not necessary. -The option button for example gives the ability to change the tables layout. -The user can also request to calculate percentages, column-percentages or no percentages at all. -Another option,for this screen only, is the possibility to perform a Chi-square test to search for a relation between the variables. -A feature, which is available in every program input screen are the selection buttons (on the top right of the screen). These enable the user to make a supplementary selection of the data. For example, the policeman can request only those robbery's happened between 8 '0 clock in the evening and 7 o'clock in the morning. Also combinations of conditions can be used. To construct the selection criteria, the police officer is given a window. No knowledge of SAS is necessary because he can build it on a menu-driven way. Figure 7 is then then the result of such a tabular report. The column variable is the week/week-end -variable (did the incident took place during the week or during the week-end). The rows are the type of offences. On the top lines of the report, a brief description of the data we are working with is repeated. 305

;Fr---='=-'""""h"--"'=-'-- Figure 7 : tabular report on the output screen Summary In this short presentation SAS-prooves the ability to handle big quantities of data in a complex way through the use

10 ;Fr---='=-'""""h"--"'=-'-- Figure 7 : tabular report on the output screen Summary In this short presentation SAS-prooves the ability to handle big quantities of data in a complex way through the use of user-friendly SASIAF applications. The system developed by M.Beullens contains a new unique registration-system of sensible data in crime matters, and provides a statistical crime-analyses with different visual presentation possibilities for the local police forces. In Belgium, this application offers a new uniform registration system used by the three big police forces ( police officers), which can -for the first time in history- give an overview of crime rates in our country. Acknowledgements The author wishes to thank the promotors of this Universitary research Prof. dr. Vuylsteke (University Center of Statistics - K. U. Leuven) and Prof. Van Kerckvoorde (Faculty of Law - criminology- K.U. Leuven) and the national coordination at the Ministery of Internal Affairs; P.Ponsaers (promotor) and E. Devroe (scientific researcher). References VAN KERCKVOORDE J., PUT V., THYS J., WOUTERS M. m.m.v. DEVROE E., PUTTAERT M.,PoIitieel geregistreerde criminaliteit, een instrument voor statistische registratie van door de reguliere polities geyerbaliseerde criminaliteit, Vanden Broele Brugge SAS INSTITUTE INC., SAS/AF Software; Usage and reference. Version 6. First edition, Cary NC, USA. SAS@,SAS/FSp@ and SAS/AF@ are registred trademarks of SAS Institute Inc., Cary NC, USA. 306

SIDRA INTERSECTION 8.0 UPDATE HISTORY

SIDRA INTERSECTION 8.0 UPDATE HISTORY Akcelik & Associates Pty Ltd PO Box 1075G, Greythorn, Vic 3104 AUSTRALIA ABN 79 088 889 687 For all technical support, sales support and general enquiries: support.sidrasolutions.com SIDRA INTERSECTION