Roads with no speed limit data in OSM have very high inferred speed limits.

[ratbopper]

If OSM does not have speed limit information, a speed limit is inferred.

• If the way ID is tagged as highway=primary, we infer its speed limit to be 50MPH.
• If the way ID is tagged as highway=secondary, we infer its speed limit to be 35MPH.

One example of this phenomenon is Massachusetts Ave in Cambridge between Putnam St. and Bigelow St. We infer a speed limit of 50 MPH here.

We also infer a speed limit of 35 MPH at Broadway and Prospect in Cambridge (two roads that are generally LTS 3), which bumps the intersection's LTS to 4.

@sckilcoyne proposed that if OSM does not have speed limit data associated with a way, we should use the default speed limit of the city that the way is located in.

This issue is related to #3, which is a general issue related to improving LTS calculations.

[sckilcoyne]

FWIW, the number we actually care about is prevailing speed, which I think means 85th percentile speed. Where given, we are using speed limit to approximate that value. For LTS, our real ideal is not knowing the speed limit of a given street, but actually measuring how fast people actually drive on the street.

Talking to Boston city staff, they do similar as us, using speed limit as a proxy for prevailing speed when they make their own stress map, but it isn't perfect.

I think the assumed prevailing speeds based on classification of streets could be refined. There might also be something where we could take into account both the speed limit (or assumed city-wide speed limit) and the road classification and average those numbers or something to come up with a hopefully better estimate of prevailing speed.