How we’re using machine learning at Transit to improve Real-Time bus predictions

I use this app. The location-based bus predictions are often accurate unless there is a significant amount of traffic. This is a problem during rush hour every week day. The bus may be traveling slower than walking speed, hitting every traffic light multiple times, and you can wait 15 minutes while the app predicts 1 minute.

I've been using this app for several months without knowing what the "Go" feature was for. I guess I could turn it on and maybe help out other riders. Although if my phone is set to only get tower-level accuracy I don't know how much it would help.

Anyway great app and I look forward to the improvement when this rolls out to my city.

What is STM? I was looking for an explanation of this acronym and couldn't find it.

At least explain or define what STM is, as this left me feeling confused.