How many recorded crimes occurred definitively according to the published data on police.uk in any given month or time period?
You may be shocked and disappointed to know that no one actually truly knows. And that goes for even our good selves, the only aggregators of the crime data. We can add up the numbers of course, but that’s not actually enough if all the data is not collected the same way in a method that is discernible. Incredible really, the government proclaims falling crime but they can’t actually keep count of what they’re publishing so how do they even know?
Were the Home Office so minded, this could be quite an easy problem to solve. Unfortunately, they are not yet and until it is resolved, it is impossible to gauge exactly how many crimes across all 43 Forces should be separately recorded each month !
Allow me to explain. The root problem is that there is no common standard and no data governance from the Home Office over the 43 Police Forces. There are 2 main sources of raw data – the neighbourhood and street csv files published here. There’s also a a crimes with no location by month and Police force string on the api. It has been said to us that crimes with no location are included within the neighbourhood files. Except as this spreadsheet shows, this is not true. They are actually all of the following;
i) included in the street crime file and labelled no location
ii) not included at all anywhere
iii) included in an undeclared basis within a Police Station within an undeclared neighbourhood
iv) only included in the api
v) some combination of the above
So the opportunities for double-counting and omissions are legion. I’m just updating a spreadsheet which I’ll be posting up here shortly which I hope will inform discussion.