Investigate directional bias in BT Data using Probe data

[radumas]

For our evaluation of new Bluetooth routes for the King Street Transit Pilot https://github.com/CityofToronto/bdit_king_pilot_dashboard/issues/41 we discovered a directional bias in observations on new Bluetooth routes.

This could lead to a bias in speed observations on those routes, but maybe those are reflective of conditions on the ground? The full analysis can be found here

• [x] Look at Bluetooth routes for which we have already coded in the matching set of HERE segments to see if they have a directional mismatch in observations and speed.
• [x] If so, is there a similar bias in the HERE data?
• [x] If not, find a few of the more obvious Bluetooth routes with disparities from the analysis above. Code in their matching HERE routes.
• [x] Is there a similar directional bias in the HERE data?

[radumas]

I created a data_validation folder for Bluetooth in master. can you clean up your folder to contain 1 main notebook at the route and then create a sub-folder for the other notebook and sql files named around the directional bias we were looking for.

And also add some info to the README

[jovenc]

Is it better now?