wrong public transport segments

datamapio / posmo_one_testing

Public Testing Repo

0 stars 0 forks source link

wrong public transport segments #28

Open thohug opened 3 years ago

thohug commented 3 years ago

Setting Zurich, Switzerland, Android

wifi-fence: 1 do_not_track: 0 wifi_over_gps: 1 data_over_wifi_only: 1

Issue Random public transport segment at days beginning (no PT was used that day, 5th decembre), only bike. And even if it was done by bike it is pointing in the wrong direction.

grafik

Another example of a correct segment pointing in wrong direction (28th nov):

grafik

rogerfischer commented 3 years ago

Good catch. Thank you! @djr4 We use course for the direction only for now. Correct? Looks like this is not always reliable. Maybe we should use a direction vector from A to B instead.

Problem 1) Did you check the raw data? What mode was used? Can you upload a picture?

djr4 commented 3 years ago

Actually implementation is not considering course but uses subsampling of gps points and calculating direction of that vector.

There is also some definition of interval when to calculate / display points.

5 consecutive gps points A1, ... A5 direction is calculated as a direction of the vector A1,A5

I'm not sure what happened here, specially for bike (guess would be that A1, A5 moment is chosen where there is actually bike noise but it should be looked in raw data

rogerfischer commented 3 years ago

@djr4 And taking the vector from static to static / next segment?

djr4 commented 3 years ago

I can't really do that, would be even less correct than now. For example for walking or modes that could have a lot of turns this would show really bad results. Would be interesting to see if this is wrong due to being mostly static at moments where this calculation happened.

thohug commented 3 years ago

Important to note that at least the first exmaple the segment is completely off (timewise mainly, starting at 0:00 until 8:37); actually I started that trip at approx 8.37 by bike.

Can't check raw data - it seems there has been a change and they're not shown anymore on that page :)

rogerfischer commented 3 years ago

@thohug Ok. Thanks. We look into it.

As for raw data, sorry my bad. Yes we secured this (it was, if you knew the id, open before). We check how we can make it available for you again. Will let you know.

djr4 commented 3 years ago

@thohug Ok. Thanks. We look into it.

As for raw data, sorry my bad. Yes we secured this (it was, if you knew the id open before). We check how we can make it available for you again. Will let you know.

This now again available for private usage.

thohug commented 3 years ago

New case of changed directions. Problem seems especially with trains, I haven't yet had any bus routes affected.

thohug commented 3 years ago

Problem with wrong directions on train segments keeps persisting (two trips on top of each other):

grafik

@djr4