Insights from Most Dangerous Districts
The above Tableau dashboard, Most Dangerous Districts, provides an interactive way to explore the criminal incidents that have occurred throughout Washington, D.C. The above map is divided by each of the 7 districts of the city using different colors. There’s a filter on the right where a user can select different years to see how the density of the incidents per district changes. As the incidents change on the map, a further breakdown below can be referred to as well.
The stacked bar chart on the left provides an overall reference while switching through the different years. The table on the right breaks down the number of incidents by time of the day. For example, and by default, 2019 is selected and the District 2 of the city appears to have the highest criminal incidents for the year while occurring during the evening shift. Based on the district breakdown, these concentrations of criminal incidents are consistent with the hot spots identified in the cluster analysis. Districts 2, 3, and maybe 4 could be considered as high priority areas for law enforcement to maintain a presence of to be watchful of future criminal incidents.
Offense Forecasting
Predictive policing refers to the usage of mathematical, predictive analytics, and different analytical strategies in criminal analysis to identify unlawful activity. Forecasting has been a popular methodology to provide law enforcement leaders a foundation how to best prepare resources for potential fluctuations of the rate of crime by geographic location.
The clustering analysis helps analysts know where to focus resources. By pinpointing the number of incidents by area and learning how those values have changed overtime, forecasting can be applied to predict the potential number of new incidents. Factors such as seasonality and current trends or patterns are also assumed to be consistent going forward. Since the recent pandemic has already greatly reduced the number of incidents as we already have seen in our analysis, forecasting based on the current actual number of incidents and new trends will should be interesting.
View Offense Forecasting