What to do Once You Have Found Them

The provisional identification of an area as a hot spot is a first step in deciding where to look for the factors which drive crime. Sophisticated spatial analysis makes that first step quicker and is worthwhile. No-one should believe, in crime audits or elsewhere, that it represents more than the first step. Dr Snow cracked the cholera epidemic because he thought he should map deaths against water sources, having the hunch that the disease was water-borne. Without that, he would just have had a series of dots, and in due course a lot more dead people.

Suggested below are some tactics for determining the meaning out of areas provisionally designated as hot spots:

• Stable hot spots are those areas with a consistently high crime level relative to other areas. These areas have long-term problems and will probably continue to experience high crime unless focused change occurs.

• Unstable hot spots describe areas that sometimes have a high crime level relative to other areas. Within this group are:

  • flare-ups – areas which experience high crime for short periods, at times difficult to predict;

  • intermittent – areas which oscillate between high and low crime ; and

  • seasonal – areas which display high crime through certain times of the year.

To address this point, it may be helpful to calculate the variability of the problem over time. Partitioning the time window for a designated hot spot into smaller sub periods and seeing how much they vary. This is best calculated as the standard deviation of sub period totals. Standard deviation is one of the functions available on software such as Microsoft Excel. Then calculate the coefficient of variability (standard deviation/mean). Hot spots with high coefficients of variability are those where the problem fluctuates greatly from sub period to sub period. Those with low coefficients of variability are those where the problem remains pretty constant over time.

Where there is a high coefficient of variability, can this be linked to anything else? Was it seasonal, at times when there was a high student population, when the local soccer team is at home, when the circus comes to town? Figures 1 shows the number of robberies in one area, then subdivided into those involving bikes and others. Without that subdivision, it would appear that robbery generally peaks in the summer, and remains high throughout autumn. It will be seen that the robbery concentration is thus clarified by further analysis. This will be time-consuming because of manual inspection of records. Hot spotting should be seen as a crude process of limiting the scale of this necessary further analysis.

A final and fundamental point involves the distinction of hot spots yielded by a few people or households being victimised often, and many people or households being victimised once. What one does by way of prevention will depend on which of these applies. Some crime maps will colour code repeats, but the accuracy of these will be limited by the adequacy of the recording system generating the maps to identify repeats. The tendency of the same people to be assaulted in different places will not appear at all. As with the robbery example, this kind of analysis typically involves time-consuming manual editing of records.

Back to Hotspots