Public Safety Analytics and Academia

Many public safety organizations are keenly interested in deriving value from the massive amounts of data currently available to them. In situations were analytic staff are not available to work with this data, or when it makes sense to analyze data in several different ways, organizations are partnering with educational institutions.

The many ways in which such partnerships can be created, however, are not always clear. In this paper, I’ll describe a few of the collaborative projects the Ottawa Police Services (OPS) and the Telfer School of Management’s (TSOM) MBA program have conducted to outline viable partnership approaches that provide benefits to students, managers and to the community at large.

There are essentially three key steps in the partnering process: build bridges; define the scope and objectives of projects; and deploy student and professor resources. One other factor to consider is the results that accrue from these types of collaborations because it is results that enable long-run sustainability of the partnership.

Figure 1: Severe Accident Decision Tree

Build Bridges
In many universities, systematic processes for building bridges between researchers and external organizations are lacking. Partnering, therefore, is often the result of individual relationships among researchers and representatives within the partner organizations. The OPS studies done by TSOM MBA students were initiated in much this way: personal relationships between researchers at the university and representatives at the OPS led to an initial discussion about the types of projects that would be of value. A series of conversations with a variety of professors ensued, during which the OPS defined some of the projects that would be of interest and the professors considered how they might become engaged.

Define the Scope
One of the key challenges for university-industry collaboration is the need for projects to fit within an academic calendar. For students in the MBA program, this means that a project could be done as a class assignment or as a research project outside of class. Within the TSOM MBA program, students can work on “Directed Readings”: research projects that span 6 or 12 weeks enabling the student to explore areas of interest that might not be offered in courses. This is the approach taken with the OPS collaboration. Three groups of students conducted different projects. The first focused on defining a “Quality of Life” indicator to allow the OPS to think though how they assess the community impact of the work they do. The second explored the impact of preventative measures taken by the OPS to reduce vehicle break-ins. The third explored several years of accident data to develop algorithms that predict the likelihood of severe accidents occurring at different intersections.

Deploy Resources
Students need to receive a grade for their projects. Accordingly, each project must be supervised by a professor and, this can be another challenge for setting up the collaboration. Typically, professors who have an interest in the study itself will work with the student therefore deployment involves the matching of a professor, a student and a representative from the industry partner. In the case of the TSOM, several professors were interested in the topic area therefore finding a match was not difficult. On the OPS side, the team working on analytics was also readily available, a factor which greatly facilitated the collaboration.

Results
Results come in various forms. For the OPS, the Quality of Life project produced a framework and a set of measures that managers could consider using to evaluate long-term impact of their community involvement activities. Many other measures are already being used, so the value to the OPS in this project resulted from the exploration of different ways in which quality of life measurements (such as reduced fear and anxiety for example) could be deployed to better evaluate the impact of community policing efforts.

The project on preventing vehicle break-ins identified the fact that there were several gaps in the data needed to track the impact of the preventative initiative. The students doing the work were able to extrapolate a number of data points to demonstrate how the initiative could be evaluated.

The traffic project defined a decision tree (among other algorithms) that sought to clarify the likelihood of certain categories of drivers being involved in severe accidents. An abbreviated version of the tree is shown in Figure 1. Tracking upwards from the “Severe” accident box, one can see that a certain combination of “Date” (note that date in this case in in Excel format), “Driver Age” and “DayOfTheWeek” tend to lead to severe accidents. The findings are not conclusive, but they do provide a pathway for further exploration, something a new group of students in the 2016 MBA cohort are in the process of doing.

Conclusions
These projects provide exceptional “real-life” assignments for students and the findings generated food for thought for the OPS team. They also generate research products that find their way into some of the courses taught by the TSOM professors. Overall, although these collaborations require a bit of time to set up, they represent a winning combination for all parties involved.

–––
Gregory Richards, MBA, Ph.D, FCMC Director, MBA Program, Telfer School of Management