Hot Spot Map
Click on the links below to download a copy of the charts from this video. Once the new window opens right-click on the image and select "Save image as...".
Traditional “dot maps” that depict each crime event in a region using a dot or pin icon suffer from a number of limitations. For example, multiple incidents at a single location show up as a single “dot” and maps with too many dots can be visually overwhelming. To address this crime analysts sometimes create “hotspot” maps using something called kernel density estimation. This type of map colors each region within a city based on the number of crimes in that specific area and nearby locations. Density maps make it easier to pinpoint areas with higher concentrations of criminal events. The maps presented here document the density of street robbery from 1995 to 2010.
To create the maps were first plotted all of the incidents for 1995. Based on the distribution of events that year we created cutoffs for coloring the map based on the overall average density throughout the city. Areas shown in white had either no street robberies that year or they had a density that was below the average for the city. Areas in yellow had a density ranging from the average to just below twice the average. The density of crimes in areas colored orange ranged from twice the average to just below three times the average. Finally, the areas colored red had a density that was three or more times the average for the city.
The cut-points created using data for 1995 were then applied to all of the subsequent years. This procedure allows for the ensuing maps to reflect increases and decreases that occur in crime counts and it also allows the viewer to identify changes in the location of hotspots over time.
It is important to remember that this sort of mapping does not control for population differences across regions of the city. An area may show up as “red” or a hotspot largely because there are a high number of people living in or travelling to that area, and hence, there are more targets available for offenders. A given person’s actual risk for victimization in such an area may actually be quite low despite the high number of offenses happening there.
Another limitation of kernel density maps is that crime in border regions may be artificially deflated. Recall that density maps are colored based on the count of crimes in a specific area and adjacent locations. In the case of our city’s borders, like Gresham to the East and Beaverton to the West, the adjacent data are not available because they are maintained by another agency.
Turning now to what these maps tell us about street robbery in Portland, you may have noticed from the video that the “hotspots” appear to be shrinking over time. Between 1995 and 2010 street robbery decreased by 60%. This decrease has not been distributed equally: For instance, Northeast Portland (just west of I-5 on the map) has experienced more pronounced decreases in street robbery. Other areas like downtown Portland have not experienced as dramatic a decline. The maps also show some movement of hotspots to the East with the emergence of distinct pockets in regions bordering Gresham.