Let me explain the features of the data set that are relevant to this post. An incident is created from information from a member of the public (or on occasions from police officer information). This is coded to the type of incident. There are about 100 codes that can be used in three fields. When the incident is completed the same codes are used in three fields to result or class the incident.
I am interested in just one of the codes "1" which is "violence against person". I am interested in this analysis of the demand profile so it would be logical to look at the type fields based on what the initial information is. I decided however to base my analysis on the class fields because I think this more accurately reflects the true nature of the incident.
So the incidents I have selected have the number "1" in one of the class fields. This gives me a rich data set of incidents to analyse, and one which I suspect is more reflective of events of violence than the recorded crime data set.
On reviewing my data I noticed that these figures are slightly inaccurate due to a small amount of missing data. Nothing to change the conclusions made. One of the problems of publishing as you go along.
ReplyDelete