Overview
When the user runs the model, they are presented with a GUI (Figure 3) that allows them to input a series of variables and files. These are inputs that the user will be able to change to ensure that their model best suits their city. The city variable is the city that the model will be run in. This ensures that all addresses end up in the correct city (for example, if an address is 800 Main St., the city needs to be placed onto the address, or this may end up in another city with 800 Main St.). The Officer Count variable represents the number of officers to be placed in the model. The default is 100, but this can be changed, depending on police force size. The Default Weight variable has the weight for an incident that is not documented in the Crime Weight File. In the example shown in Figure 3, if the incident type "robbery" was not given a weight in the file weights.txt, the weight for a robbery would be 50. This can be changed based on what the most prominent crime type in the city is. When running our model, the default weight is set to 50, signifying that all undefined weights will be defaulted to the weight of a violent crime. The variable Grid Size indicates how large (in meters) the grid to place officers in should be. The variable Max # of Officers per Grid indicates how many officers should be placed to patrol in one grid area. The maximum value this variable can be is 10.
The files inputs allow the user to add their own files as the input and output files. The Data File is the tab-delimited data file where the crime data is placed, containing address, time, and incident type. The Crime Weight File is an input file where a crime type is accompanied by a weight that the crime type should be given. Most crime types should be identified here, but if one is missed, the Default Weight will default the weight of that particular crime to its value. The Predicted Crime Output File is the output data file containing a list of latitudes and longitudes and the weight each one should have. This file is used to map the crime data as well as calculate where the locations of police officers should be. The Police Location Output File is the file that maps the latitudes and longitudes of the calculated police locations. The default value of both output files is the necessary name scheme for them when they are uploaded to the web map display. Once the variables and files have been input, the forecasting program runs, and a progress bar appears to show the user how long is remaining in the place initialization. This can sometimes take a long amount of time, so the progress bar is able to tell the user how much of the data file has been initialized. Once the forecasting program has exported its file, the officer placement program runs using the file from the forecasting program. The officer placement program will then export a file that contains the best positioning of officers. These two files can then be uploaded to our website in order to visualize the data as a heatmap, with points overlaid for where officers should be placed for maximum efficiency (Figure 4).
Figure 4: An example heatmap with officer points overlaid in places where they would be most
efficient. Numbers indicate how many officers should be positioned in an area.