I am carrying out a research on how complex city road networks get and I need to get accurate map information.
I started with openstreetmaps however I ran into a couple of issues: some of the dumps contain duplicate information – for example two ways which have the same name tag have distinct way ids and this skews my representation.
Moreover, when I converted their data into a graph, the graph was not connected, infact about 30% of the roads are not connected to the giant component. Could this be a mistake in my representation or something to do with openstreetmaps?
I have also tried dumps from the US geological survey however roads are split into segments without enough information to recombine them again. This makes it impossible to get full road information.
Can you help with road network data suggestions?
Best Answer
I'm sure you're aware of the nature of Open Street Map - it's a user collected and edited resource. If you're going to use this kind of information it would be largely down to you to fix these - and in the spirit of things update the information!
If you want a relatively clean network for analysis straight away, I'd suggest TIGER.
... but I think you'd have to attribute information such as speed limits and the like.
Otherwise, it's going to cost money - have you got a budget?
UPDATE Just downloaded a Tiger file for the roads - you have to extract it on a county by county basis (so will have to join layers together if crossing county boarders). But the file is coded by road classification, so you could attribute these with median speed limits quite easily).
MORE UPDATE This kind of question appears to have been answered here: US Tiger or Open Street Map Route Data: One Way Street & Intersection Data