As discussed here it'd be good to have a way of testing how 'good' our map matching is by comparing it with known results. (I.e. our map-matching on a GPX track gives a certain route, and how close does that align to the actual known route taken?)

Considerations/comments (to direct a PR):

• where do we get data of known routes? @stefanholder pointed out one example - do we want others? Do we want users to be able to test their own data sets?
• we should see how map matching performs not only with the given GPX track, but also as we continually down-sample it.
• how do we compare with the known route? Compare total distances of each? Or the proportion of the distance which they share common edges on?
• how will we add this functionality? I think it is very similar to the Measurement suite (Performance measurement suite #73) in how it'll be used (and we could, I guess, put it in there if we wanted). Or we could build it into unit tests somehow (e.g. require that the match is sufficiently high.)
• do we want to store historical trends in accuracy (as opposed to e.g. just comparing with the last commit) - and if so, how?