Rochester News

 

RPD's New Patrol Districts

 

The RPD released its new patrol districts. Below is their official release and the new maps...

 

Data Analysis Overview

The intent of the data analysis is to create a baseline to make data-driven process decisions that allow flexibility of staffing and beat models for continual evaluation and adjustment. The Rochester Police Department (RPD) recognizes that comprehensive data analysis is a key component to informed decision making but it must be balanced with experiential knowledge, existing constraints, and practical considerations.

Workload model

The purpose of establishing a workload model is to create a basis for the evaluation of demand for service on the patrol division. The primary demand on patrol officers is non-discretionary calls for service, calls placed by citizens that require a police response. Calls for service data are commonly assessed in raw counts, however, to effectively analyze demand other factors must be considered. To address the need for a comprehensive analysis of demand, RPD built a six-variable weighted workload model.

Data Analysis Overview

The intent of the data analysis is to create a baseline to make data-driven process decisions that allow flexibility of staffing and beat models for continual evaluation and adjustment. The Rochester Police Department (RPD) recognizes that comprehensive data analysis is a key component to informed decision making but it must be balanced with experiential knowledge, existing constraints, and practical considerations.

 

Total Hours Worked (75%)

To calculate total hours worked, RPD analyzed 5 years of calls for service data or roughly 2.5 million data points. The pattern of non-discretionary calls for service responded to by RPD’s patrol division has remained consistent over many years, and is highly predictable. Five years of data is a sufficient sample to draw conclusions from and also broad enough to account for the minimal year-to-year variance. Proactive calls for service were eliminated and metrics (e.g. average call length, average number of assisting cars, etc.) were established for each of the remaining 119 unique call types. An upper-bound of average call length was calculated for each call type, defined as 1 standard deviation above the mean.

The formula used to give each call type a single numerical value for total hours worked was: Hours Worked = Upper-bound + (Avg. Call Length x Avg. Number of Assisting Cars)

 

Calls for Service (10%)

Two years of non-discretionary calls for service were spatially referenced as a baseline for geographic comparisons. Proactive calls were again eliminated.

Average Drive Time (7%)

Average length in minutes from dispatch time to arrival time was calculated by weighting two years of  non-discretionary calls for service data for each of the current 22 Police Service Areas (PSA). Emergency calls were given a higher weight than Non-Emergency calls.

Population Density (4%)

2010 US Census Block level data was spatially referenced as a baseline for geographic comparisons.

Area (2%)

Total area in square miles was calculated using 2013 city borders as a baseline for geographic comparisons.

Street Segments (2%)

Total length of street segments in feet was calculated using the Monroe County Centerlines shapefile as a baseline for geographic comparisons.

 

The six variables were then uploaded for additional analysis into a Geographic Information System (GIS), specifically ESRI’s Districting Tool. Using this tool, the City was broken into 250 feet by 250 feet grids, each containing a numerical value based on the workload model weighting. This provided a single value for each grid cell representative of its workload proportional to the entire city. As a proof of concept, RPD ran multiple iterations of this weighted model using historic data against the current Divisional boundaries and quasi-operational Quadrant boundaries for comparison purposes. The model was closely aligned with the Department’s understanding of the current workload balance. At this point RPD is confident that the model is a useful tool to gauge workload in a geographic context and flexible enough to allow modification when necessary. RPD will also continue to test the validity of the model.

 

Future Analysis

Staffing

RPD has begun a data-driven (e.g.platoons, work wheel, vacancy rate, show rates, etc.) approach to analyzing the projections for the organizational staffing in the recommended 5 section model.

Car-Beats

RPD will utilize the workload model in combination with qualitative feedback from its officers and the community to develop new, neighborhood-based patrol beats within the new section boundaries.

Evaluation

 

RPD will establish key performance indicators for ongoing evaluation of reorganization.

 

 

 

 

 

 

 

More Articles