Huh… I am no expert (but some here are: for example @mstn, @elf_pavlik, @napo and @simonecortesi). I have never studied this in detail. But:
- I don’t understand what you mean by “data analysis cannot be implemented in OSM”.
- As for computing accessible itineraries; of course you are right that this is not done in OSM itself. The whole idea of OSM (or Google Maps), as far as I understand it, is that it provides the data for other people to build applications on top of. The OSM wiki has several routing examples optimised by different needs (for example this one for cyclists).
- Any data you collect that do not meet the OSM stringent requirements, like the 10 simple questions, can be stored on your own server and “mixed” with map data on the fly via API calls.
So, it seems that (in theory) you could build an added value application on top of OSM as of Google. The difference is that Google data are not plastic the way OSM data are: any edit to the Google database you have to ask for, then wait and hope, whereas in OSM you can just go ahead and make the edit yourself. This has big implications in terms of building community and awareness. For example, you can organise mapping parties.
Besides this technical issue, you might enjoy the work that the City of Milan is doing in a similar vein: trying to build cheap, simple solutions for shopkeepers to make their shops wheelchair accessible. This initiative is spearheaded by @rossana_torri and @alessandro_contini; @federico_monaco is observing it closely. The project is called Open Rampette.