It's likely to be in that, as the police were called
[snip]
It is however a bit wierd what data is on there, as lots of it seems to be in small streets that wouldn't affect TFL's operations at all, so why do they have the data?
It's not TfL's data--it's collected by the police and therefore for all crashes to which they are called (bar the odd administrative error). TfL have a unit that works on analysing and publishing these stats. It's just a bit easier to collate it centrally than to let every London local authority do it separately. TfL are, of course, particularly interested in crashes on the TLRN.
At least some of the plots in side streets would appear to be mistakes, too. I suspect that many are not correctly plotted at junctions (where the vast majority of crashes occur, I think the latest figure was 84%), but at the location used in the mapping for a link (one of the intersecting streets), which often seems to be halfway along the street. I don't know much about on-line mapping, though.
Police data is known to be highly incomplete owing to the level of under-reporting. It is thought that only fatal crashes don't attract under-reporting, with quite a lot even with serious crashes, and pervasive under-reporting for slight injuries.
It's not TfL's data--it's collected by the police and therefore for all crashes to which they are called (bar the odd administrative error). TfL have a unit that works on analysing and publishing these stats. It's just a bit easier to collate it centrally than to let every London local authority do it separately. TfL are, of course, particularly interested in crashes on the TLRN.
At least some of the plots in side streets would appear to be mistakes, too. I suspect that many are not correctly plotted at junctions (where the vast majority of crashes occur, I think the latest figure was 84%), but at the location used in the mapping for a link (one of the intersecting streets), which often seems to be halfway along the street. I don't know much about on-line mapping, though.
Police data is known to be highly incomplete owing to the level of under-reporting. It is thought that only fatal crashes don't attract under-reporting, with quite a lot even with serious crashes, and pervasive under-reporting for slight injuries.