I'm running ArcMap 10.3.

I'm looking to analyse road collision data provided by a client. Each collision is recorded as a point, with a severity allocated to the accident. For example fatal accidents have a score of '3' whereas minor shunts and bumps have a score of 1. My client is interested in one kilometer stretches of roads which have a combined 'score' of more than 7 for the accidents therein.

In the past I've simply used a 'Polyline Dicer' tool to split a model of the road network into kilometer lengths. However this is a 'dumb' process, and it is possible that stretches of the road network that meet the above criteria will be missed.

I've considered cluster analysis but 1) there seems to be a slightly wooly element to how you set the parameter and 2) cluster analysis outputs won't flag up sections of the road network. For example I can identify the roads within a cluster hotspot, but I will also erroneously flag up side streets and other unrelated road lengths.

I'm aware of the 'SANET' toolset which would perform the sort of analysis I'm interested in. I can potentially commission another organization to carry out this work (I can't obtain a SANET license directly for my line of work), but I'm wondering if there is a deft combination of tools in ArcMap that would give me the same, or similar, results.

In my head there is a fuzzy workflow that involves step-wise buffer analysis of each accident to consider all other collisions within a 500 meter radius of the given collision, identifying those which have the same underlying road attribution (road name/type) and somehow working from there.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.