Police.uk: A Rant
As part of the big society the drive to understand local crime statistics has resulted in the police.uk website, while its aims are ambitious, it seems to be neither well designed or thought out.
Issues within the dataset first came to light namely the geo-location of reports, be it whether they were anonymised enough, to the majority of crimes being reported at police stations, therefore showing a large incidence of crime at police stations. Where is the ability to show time – this is important in crimes, if crime happens in a town centre at night, this makes sense, people drink. However if the majority of incidences of crime occur during the day, this is worrying. This all boils down to the usability and visualisation of geographic information, but as it’s the government’s first shot at making something like this so fair enough, right?
The greater crime about this is the facility of reporting the crime and the presumable steps behind the system. Currently the protests around the world in Tunisa, Eygpt use a platform called Ushahidi, also used to great effect in the oil spill and in various other instances that require citizen participation/reporting. This version looks like a poor imitation, while not retaining Ushahidi’s ease of use. Surely to visualise the crime on a first step shoehorning the police database into one that Ushahidi understands would be better?
A hue and cry from a wide plethora of people from the daily mail to estate agents (probably the same thing) were complaining regarding the effect it would be having on house prices. This on the face of it seems to be a decent point, however given the analysis done by companies such as Experian and insurance companies on the level of crime in an area, makes a moot point, surely if people care enough about buying in an area they don’t know they (or their surveyor) will do the research, it is a big investment after all.
One point of note was the way information regarding local policing has been laid out, showing who is in your local police team with their contact details etc, though it barely a leg to stand on.
There was a Michael McIntyre gag about google streetview in that given the wealth of information and things to be seen, the first thing people look at is their own house, something that could be achieved just by going outside and facing your own house. So when Jonny public finds out about a magical service that can see what is going on in their area, the inquisitive England are going to log on in their millions, and be outraged if it works “There is crime in my area, why aren’t there more police” or if it doesn’t “It’s my tax, that I’m paying, why isn’t this working?”. Unfortunately pleasing everyone is difficult, and it wasn’t made easy with their start. Scalability was also seen to be an issue, surely a cloud service like Amazon should have been employed?
Much better examples of a crime sites exist, namely the excellent Oakland Crimespotting, especially the ability to break down and further analyse the crime in a given area, the omission in the British version of serious crime like rape and murder seems to not give a proper version of events. It may be hard to stomach however crime isn’t pretty, however it should be shown in all its blood and guts just to fulfil its remit.
The basemap also raises eyebrows. Sure Google is familiar to everybody, however we have two other sensible competitors, Ordinance Survey and OpenStreetMap. The ordinance survey on the face of it seems the better choice; It is the national mapping agency after all, we’ve already paid for it. Their products are already up to the task, and the government could have probably shoe horned it into their remit and made them do it through a matter of course. It does, however, share the disadvantages of Google and other mapping agencies, in that the data isn’t often updated in the same speed and manner as OSM.
OpenStreetMap on the other hand (in my opinion) would be the better choice. The map is updated faster than the other traditional agencies (after all Google will licence the data, that in part will come from the OS, as well as TeleAtlas and NAVTEQ). Questions about the quality of OSM have been raised in the past, however the longer it goes on the stronger OSM gets, it now finds itself in position as the defacto choice for the White House, crisis situations through Humanitarian OpenStreetMap Team (HOT) and Map Action and a whole host other applications. It’s quality isn’t in question and the produced map wouldn’t be subject to Google’s TOS with OSM just requiring attribution, then through using providers like CloudMade as a file server a custom crime map, instantly recognisable could be produced, against the Google one.
To sum this up, I’m not against the police.uk website, I’m massively for it and as a first try it wasn’t bad, however £300,000 for a Google mashup seems slightly steep. I’m unsure how much was custom written, though it looks like an Ushahidi clone, surely bolting the police dataset to an instance and fixing the geocoding – the issues should have been solved by better testing – would have been cheaper and more informative.