Measuring a Place’s Exposure to Facilities Using Geoprocessing Models: An Illustration Using Drinking Places and Crime

Abstract

The prominent role of facilities in influencing ‘why crime happens where it does’ has been widely recognized and vigorously researched. Criminological theories which focus on opportunity such as routine activity theory and crime pattern theory have provided the basis for such inquiries. Some of these investigations have targeted the role of facilities in fueling higher crime levels at places. They have usually developed facility-focused measures that quantify each facility’s influence based on the crime experienced by the places located near it. Measures are calculated only at the locations with facilities present. However, improvements in data sources and software have allowed researchers to examine the population of small units of geography rather than focusing on only those with a facility present. Thus it is now possible to quantify the cumulative effect of nearby facilities on the crime rates of geographies of such street blocks and addresses. This chapter begins by discussing the traditional methods for exploring the relationship between facilities and crime. Next, the theoretical case for more sophisticated distance and activity-level based measures is made. The critical role of geoprocessing models in automating complex analysis processes is explained and a model developed to create three different exposure measures. Data describing the locations of drinking places and street block level crime are used to illustrate how measures produced by the model can be used to explore the relationships between exposure to facilities and an outcome such as crime. The output measures from the model are evaluated using descriptive statistics and then used as independent variables in an ordinary least squares regression. Local variation in the measures is examined using a bivariate LISA to highlight areas of negative and positive spatial autocorrelation between exposure to bars and crime. The chapter concludes with a discussion of the implications of the findings and probable next steps.

Appendix: Documentation of Geoprocessing Model for Calculating Exposure to Facilities

A geoprocessing model was used to develop the cumulative measures of exposure to facilities. There are three main inputs to the model: (1) a network data set; (2) a feature class of point locations representing origins; and 3) a feature class of point locations representing destinations. The remainder of the Appendix provides a step-by-step description of the model. A snapshot of the section of the model appears first followed by a description of that step.

Step 1:

Make an OD Cost Matrix Layer – This tool makes an origin–destination (OD) matrix. The user sets the analysis properties for the matrix. This type of analysis is especially helpful when the goal is to specify the “costs of going from a set of origin locations to a set of destination locations” (Esri help 2011). This is where the user sets the impedance attribute and the threshold distance cutoff. The impedance attribute is the field is used to measure the distance or time between each origin and destination pair.

Output: OD Cost Matrix -

Step 2:

Add Locations – This tool adds locations which are network analysis objects to the network analysis layer. In this model, the first locations added are the origin locations which consist of the points representing the mid-points of street blocks (which by definition include both sides of the street between two intersections).

Step 3:

Add Locations – This tool adds locations which are network analysis objects to the network analysis layer. In this model, this second instance of Add Locations tool is adding the destination locations which consist of the points representing the type of facility (here it is drinking places).

Output: OD Cost Matrix (3)

Step 4:

Solve tool – Uses OD Cost Matrix (3) and the properties identified earlier to measure the distances between origin and destination points.

Output: Network Analyst Layer – exists in virtual memory

Step 5:

Select Data tool – Selects the Lines data element developed by the Solve tool which exists in the Network Analyst Layer stored in a geodatabase.

Output: Lines data

Step 6:

Copy Features (2) tool – Copies the features from the input feature class into a new layer in the geodatabase for further manipulation. Resides under work/scratch.gdb.

Output: outputodmatrix

The output matrix looks like this:

Notice the ‘Name’ field has the origin node number followed by a dash and then the destination node number. In order to work with these, we need to get them into two separate fields. The field to contain the origin data is called ‘UofA’ and the field to contain the destination data is called ‘DP’. The next several steps are to add the fields and then calculate their contents using the contents of the ‘Name’ field.

Step 7:

Add Field tool – Adds a field to the specified the input feature class (in this case outputodmatrix) in the geodatabase for further manipulation. The feature class resides under work/scratch.gdb/outputmatrix. The field added is called UofA (field type  =  double). The purpose of this field is to hold the five digit unique unit of analysis number.

Output: outputodmatrix (2)

Step 8:

Add Field (2) tool – Adds a field to the specified the input feature class (in this case outputodmatrix (2)) in the geodatabase for further manipulation. The feature class resides under work/scratch.gdb/outputmatrix. The field added is called DP (field type  =  double). The purpose of this field is to hold the X digit unique unit of analysis number.

Output: outputodmatrix (3)

Step 9:

Calculate Field tool – Calculates the values in a field according to an expression supplied by the user. In this case, the expression (theValue) is created from a block of Visual Basic (VB) code which calculates the left side of the ‘Name’ equal to the ‘UofA’ field added in Step 7.

theVal

theName  =  [Name]

theLoc  =  Instr(theName,”-”)

theValue  =  Left(theName,theLoc −2)

The first of code creates and sets a variable called ‘theName’ to be equal to the field called ‘[Name]’. The second line of code finds the position number of the dash in the contents of the ‘theName’ field. The third line subtracts 2 from the variable ‘theLoc’ (this was the position of the dash in the string). Output: ­outputodmatrix (4) Output: OD Cost Matrix (2)

Step 10:

Calculate Field (2) tool – Calculates the values in a field according to an expression supplied by the user. In this case, the expression (‘theValue’) is created from a block of Visual Basic (VB) code which calculates the right side of the ‘Name’ equal to the ‘DP’ field added in Step 8.

theVal

theName  =  [Name]

theSize  =  Len(theName)

theLoc  =  Instr(theName,“-”)

theValue  =  Right(theName,theSize – theLoc)

Similar to the code in Step 9, the purpose of this code is to extract the numbers representing the destination point’s unique id. The first line of code creates and sets a variable called ‘theName’ to be equal to the field called ‘[Name]’. The second line of code creates and sets a variable called ‘theSize’ equal to the total length of ‘theName’ variable. For the first record in the sample below, ‘theSize’  =  4. The third line creates and sets ‘theLoc’ variable to the position of the dash in the string (first line below ‘theLoc’  =  6). The fourth line creates and sets a variable called ‘theValue’ to be the rightmost part of the string in ‘theName’ variable equal to the difference between ‘theSize’ and ‘theLoc’ (same example , 4−6  =  −2). The Expression box tells the computer to set the value of ‘UofA’ equal to ‘theValue’.

Output: Lines (5)

After this step output matrix looks like this:

Step 11:

Make Feature Layer tool – Creates a feature layer from the lines representing the distances from each origin to each destination. This layer is temporary and will not persist after the session ends.

Output: Output Data Layer consisting of outputodmatrix_Layer

Step 12:

Add Join tool – Joins a layer or table view to another layer or table view based on a common field. In this case, the join is from the drinking places shape file to the Output Layer using the DP field. This allows the attributes attached to each drinking place to be used in the calculation of a measure.

Output: Lines (6) a composite of outputodmatrix_Layer

Step 13:

Copy Features tool – Copies the features of the input layer (in this case Lines (6)) to a new layer or feature class. The new feature class is stored in work\scratch.gdb\lines

Output: Lines (7) a composite of outputodmatrix_Layer

Step 14:

Add Field (3) tool – Adds a field to a table. In the model, this is adding the field ‘i_Exp04’ to hold the exposure values for inverse distance weighted count of drinking places in 2004. A value is calculated for each origin–destination pair within the threshold distance. In this case the influence of the drinking place is reduced from 1 based on the distance away from the origin. The farther the distance, the lower the resulting value attached to the drinking place.

Output: Lines (8) store in … work\scratch.gdb\lines

Step 15: Calculate Field (3) tool – Calculates the field ‘i_Exp04’ equal to the formula in the Expression box. In the model, the expression is 1- Sqr (( [Total_Length] /5280)) which is the formula for inverse distance weighting (IDW) of each drinking place.

Output: Lines (9) store in … work\scratch.gdb\lines

Step 16:

Add Field (4) tool – Adds a field to a table. In the model, this is adding the field ‘i_ExpSal04’ to hold the exposure values for inverse distance weighted annual sales a of drinking place in 2004. A value is calculated for each origin–destination pair within the threshold distance.

Output: Lines (10) store in … work\scratch.gdb\lines

Step 17:

Calculate Field (4) tool – Calculates a value for the field ‘i_ExpSal04’ using the expression: (1- Sqr ( [Total_Length] /5280))* [drinking_places04_num_sales]. This field holds the exposure values for inverse distance weighted annual sales of each drinking place within the threshold distance of each street block. A separate value is calculated for each origin–destination pair within the threshold distance.

Output: Lines (11) store in … work\scratch.gdb\lines

Step 18:

Summary Statistics tool – Calculates summary statistics for each field in a table. For each street block (UofA field is used as unique identifier), all the values of the ‘i_Exp04’ and the ‘i_ExpSal04’ are summed. This produces a table with one record for each street block and fields containing the cumulative values for length, ‘i_Exp04’ and ‘i_ExpSal04’.

Output: Lines Sum – a new table that is written out and stored in … work\scratch.gdb\lines_Sum