Travel Mapping

User Discussions => Other Discussion => Topic started by: mapcat on June 10, 2018, 02:46:50 pm

Title: Two bugs: route page % clinched by segment
Post by: mapcat on June 10, 2018, 02:46:50 pm
KY 560 (http://travelmapping.net/hb/index.php?u=mapcat&units=miles&r=ky.ky0560) is a route driven by five users and clinched by no one, where all users have driven the segment concurrent with KY 165 (and only that segment).

Bug #1: the waypoint table shows that 100% of users have been on the concurrent segment, and 20% have been on the other two segments, even though no one has traveled the other two segments.

Bug #2: the route stats table shows that I've traveled .22 mi of the route (3.68%), and that all five travelers have also traveled .22 mi of it (3.60%). Why the discrepancy in % traveled?

BTW the KY52 waypoint name will be corrected in my next pull request.
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on June 10, 2018, 04:44:53 pm
https://github.com/TravelMapping/Web/issues/272
Title: Re: Two bugs: route page % clinched by segment
Post by: Markkos1992 on April 20, 2019, 09:28:36 am
I am unsure if this is included with this issue, but I noticed when looking at PA 347 (http://travelmapping.net/hb/index.php?units=miles&u=markkos1992&r=pa.pa347) this morning that 14.29% of the travelers had traveled between GreRidSt and DriSt (the segment I added last night).

I could not find any record in Github showing that someone had added GreRidSt to their list files yet.
Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on April 20, 2019, 11:53:28 am
Don't think it's strictly related to this issue.
I've noticed that table can be a bit janky at times, especially in the first/last entries.
According to PA-region-traveled.tmg (https://github.com/TravelMapping/DataProcessing/issues/196), there are no travelers on PA PA347 GreRidSt DriSt.
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on December 30, 2020, 01:57:38 pm
Both situations described in this thread have changed, so I can't test them directly.  I have poked around looking for similar examples and behavior for showroute seems correct in all cases I've tried.  I'm going to close the corresponding GitHub Issue.  Please reply here if it looks like I have done that too soon and the problems described still exist.
Title: Re: Two bugs: route page % clinched by segment
Post by: Markkos1992 on December 30, 2020, 02:29:53 pm
The one question I have here on this is, "Are we intending to show percentages of all users on a route or just those that have traveled the route?" for the specific route pages.  I think it is the former at the moment.
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on December 30, 2020, 03:12:42 pm
The one question I have here on this is, "Are we intending to show percentages of all users on a route or just those that have traveled the route?" for the specific route pages.  I think it is the former at the moment.

It is the former.  Not sure what people would find more interesting.  I don't think I want to clutter showroute with both, but I could be convinced.
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on December 31, 2020, 04:06:40 pm
The one question I have here on this is, "Are we intending to show percentages of all users on a route or just those that have traveled the route?" for the specific route pages.  I think it is the former at the moment.

It is the former.  Not sure what people would find more interesting.  I don't think I want to clutter showroute with both, but I could be convinced.

I'd personally prefer to see the percentages of users who have traveled the route, rather than the percentage of all TM users. That info seems more useful to me. And besides, it has been that way ever since I first became involved with this project.

You have piqued my curiosity though: Is there any segment on any route that 100% of all users have traveled on? I'd be really surprised if there was.
Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on December 31, 2020, 05:08:18 pm
Huh? We see both percentages.

Quote from: https://travelmapping.net/hb/showroute.php?r=me.me005
3 (1.02%)

3.61% of drivers
Title: Re: Two bugs: route page % clinched by segment
Post by: Markkos1992 on December 31, 2020, 05:31:44 pm
Huh? We see both percentages.

Quote from: https://travelmapping.net/hb/showroute.php?r=me.me005
3 (1.02%)

3.61% of drivers

I think I can make myself more clear now.  I think I was referring to the segments themselves. 

My current understanding is that 5.78% of all drivers have driven the segment of ME 5 between ME9 and ME98.  I wonder if it is more useful for the segment percentages to be just for those that traveled the route instead of all travelers.

EDIT:  Hovering over the "%" by the waypoint segments show the following, "Percent of people who have driven this route who have driven the segment starting at this point."
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on December 31, 2020, 06:18:46 pm
My current understanding is that 5.78% of all drivers have driven the segment of ME 5 between ME9 and ME98.  I wonder if it is more useful for the segment percentages to be just for those that traveled the route instead of all travelers.

EDIT:  Hovering over the "%" by the waypoint segments show the following, "Percent of people who have driven this route who have driven the segment starting at this point."

Clearly I need to fix either the hover text to match the values it's showing, or fix the values it's showing to match the hover text.  I think I'll fix the values, but I'll give more people a chance to chime in before changing it.
Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on December 31, 2020, 07:46:54 pm
more useful for the segment percentages to be just for those that traveled the route instead of all travelers.
This.
Yes, fix the values.
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on April 27, 2021, 02:52:39 am
more useful for the segment percentages to be just for those that traveled the route instead of all travelers.
This.
Yes, fix the values.

Would really love to see this be fixed soon... Thanks!

The percentages used to represent the percentage of people who have traveled the route; it was only recently that it got changed to all users in TM. I really prefer it the way it was before because showing the percentage of all TM users is less useful IMO.
Title: Re: Two bugs: route page % clinched by segment
Post by: Markkos1992 on April 27, 2021, 09:01:40 am
I agree with compdude.  There is no reason for this to track percentages for all users because someone like michih, who has a slim chance of traveling most routes in the US, gets tracked if doing all users just like all of us in the US.
Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on April 27, 2021, 09:17:12 am
Thirded.
Title: Re: Two bugs: route page % clinched by segment
Post by: si404 on April 27, 2021, 10:00:38 am
I agree with compdude.  There is no reason for this to track percentages for all users because someone like michih, who has a slim chance of traveling most routes in the US, gets tracked if doing all users just like all of us in the US.
And vice versa.

33 users have travelled this (one of the highest in Europe) (https://travelmapping.net/hb/showroute.php?r=eng.m004), which is over 10%, but the highest segment is 8.8%. I'm not doing the maths to work out how many that is (27 via a different route to find out).

Arguably a number would be better than a percentage for the segment totals.
Title: Re: Two bugs: route page % clinched by segment
Post by: mapcat on April 27, 2021, 03:47:21 pm
Arguably a number would be better than a percentage for the segment totals.

Agreed. Plus, this would be consistent with how the data appear in mapview.
Title: Re: Two bugs: route page % clinched by segment
Post by: andrepoiy on April 27, 2021, 04:26:22 pm
I agree. I thought that it was strange that it was only like 1% when I was the only one that travelled on a certain route.
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on April 27, 2021, 05:17:12 pm
Arguably a number would be better than a percentage for the segment totals.

Agreed. Plus, this would be consistent with how the data appear in mapview.

Yeah, I actually like the idea of having a number rather than a percentage.
Title: Re: Two bugs: route page % clinched by segment
Post by: vdeane on May 01, 2021, 06:49:09 pm
What does having only a number give us that percent of users who traveled on a given route doesn't?  I would think the percentage would be more useful - it just needs to be put back on the metric that makes sense.  If you want a number, then maybe list both, but I know I'm more interested in the percentage.  I don't really care about the raw number.
Title: Re: Two bugs: route page % clinched by segment
Post by: michih on May 02, 2021, 02:30:55 am
What does having only a number give us that percent of users who traveled on a given route doesn't?

A number is clear. No room for interpretation. With percentage you always need to know the base, i.e. what does 100% mean?
Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on May 02, 2021, 08:55:15 am
A number is clear. No room for interpretation. With percentage you always need to know the base, i.e. what does 100% mean?
That number is also displayed in the HB, so we can make the calculation.
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on May 02, 2021, 04:24:47 pm
Looking for opinions about what I have on tmstage.teresco.org (and testing of it to make sure it looks correct now).
Title: Re: Two bugs: route page % clinched by segment
Post by: Markkos1992 on May 02, 2021, 05:06:16 pm
Looks fine, but oddly enough, you see it on the connected route page for I-95 (https://tmstage.teresco.org/hb/showroute.php?u=markkos1992&r=de.i095&cr), but not on the DE I-95 page (https://tmstage.teresco.org/hb/showroute.php?u=markkos1992&r=de.i095).  I am not sure how we rectify that though.

EDIT:  Never mind, it must have been straightened out already...

Title: Re: Two bugs: route page % clinched by segment
Post by: yakra on May 02, 2021, 05:45:13 pm
Looking for opinions about what I have on tmstage.teresco.org (and testing of it to make sure it looks correct now).
I like it. Provides both the raw # & %age, without taking up any more real estate.
Title: Re: Two bugs: route page % clinched by segment
Post by: si404 on May 02, 2021, 05:57:36 pm
It's a like from me!
Title: Re: Two bugs: route page % clinched by segment
Post by: andrepoiy on May 02, 2021, 06:30:51 pm
Is it just me or are the colours still a bit off? For example, ON427 at QEW is still a yellow despite the percentage being 96%
Title: Re: Two bugs: route page % clinched by segment
Post by: Jim on May 02, 2021, 06:34:30 pm
Looks fine, but oddly enough, you see it on the connected route page for I-95 (https://tmstage.teresco.org/hb/showroute.php?u=markkos1992&r=de.i095&cr), but not on the DE I-95 page (https://tmstage.teresco.org/hb/showroute.php?u=markkos1992&r=de.i095).  I am not sure how we rectify that though.

EDIT:  Never mind, it must have been straightened out already...

Looks like I've got a race condition in there.  I wasn't happy with my solution anyway but thought it was just a little inelegant.  Turns out it's just broken.  Hope to get another look at it before too long.  Big pile of grading awaits first.
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on May 03, 2021, 01:56:30 pm
Looks good to me.
Title: Re: Two bugs: route page % clinched by segment
Post by: michih on May 03, 2021, 01:56:59 pm
The info text could be improved.

(https://forum.travelmapping.net/index.php?action=dlattach;topic=2481.0;attach=396)
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on May 26, 2021, 01:20:40 pm
It doesn't look like this has been fully fixed yet. For US10HisFre (https://travelmapping.net/hb/showroute.php?units=miles&u=compdude787&r=wa.us010hisfre), it should show 100% for the various segments since only one person (me) has traveled on this road. However, it still says 0.33% for the percentage when it really should say 100%.
Title: Re: Two bugs: route page % clinched by segment
Post by: michih on May 26, 2021, 01:57:10 pm
It doesn't look like this has been fully fixed yet. For US10HisFre (https://travelmapping.net/hb/showroute.php?units=miles&u=compdude787&r=wa.us010hisfre), it should show 100% for the various segments since only one person (me) has traveled on this road. However, it still says 0.33% for the percentage when it really should say 100%.

One traveler out of 307 TM users = 0.33% of all TM users.
Title: Re: Two bugs: route page % clinched by segment
Post by: compdude787 on May 26, 2021, 02:08:12 pm
It doesn't look like this has been fully fixed yet. For US10HisFre (https://travelmapping.net/hb/showroute.php?units=miles&u=compdude787&r=wa.us010hisfre), it should show 100% for the various segments since only one person (me) has traveled on this road. However, it still says 0.33% for the percentage when it really should say 100%.

One traveler out of 307 TM users = 0.33% of all TM users.

I thought we agreed that the percentage should be the number of people who traveled that segment out of the total amount of travelers on that route, not the total amount of TM users.
Title: Re: Two bugs: route page % clinched by segment
Post by: michih on May 26, 2021, 02:25:05 pm
It never made it to the production server:

Looking for opinions about what I have on tmstage.teresco.org (and testing of it to make sure it looks correct now).

However, https://tmstage.teresco.org/hb/showroute.php?units=miles&u=compdude787&r=wa.us010hisfre also shows 0.33%  :o
Title: Re: Two bugs: route page % clinched by segment
Post by: Bickendan on June 03, 2021, 03:11:24 am
It doesn't look like this has been fully fixed yet. For US10HisFre (https://travelmapping.net/hb/showroute.php?units=miles&u=compdude787&r=wa.us010hisfre), it should show 100% for the various segments since only one person (me) has traveled on this road. However, it still says 0.33% for the percentage when it really should say 100%.

One traveler out of 307 TM users = 0.33% of all TM users.

I thought we agreed that the percentage should be the number of people who traveled that segment out of the total amount of travelers on that route, not the total amount of TM users.
I agree with this.