search

mapourtransit / gtfs-maker

From OSM to GTFS
GNU Affero General Public License v3.0
10 stars 1 forks source link

New timetable CSV under construction #8

Closed albertocottica closed 9 years ago

albertocottica commented 9 years ago

https://docs.google.com/spreadsheets/d/1zV_WcNrjuX9ErcmBcHVU9-CW57Qwm2JpNTqOOxjic3I/edit#gid=0

So far I have input lines 2 and 3, on top of line 1 already input by Marco.

albertocottica commented 9 years ago

Line 4 done. Two problems:

  1. Line 5 has more than two endpoints. How should we report it in the data?
  2. Line 6 is a circular line. Do I simply omit the return fields?
albertocottica commented 9 years ago

Solution for circular lines (also Line 12): leave the r_start_time and r_end_time fields blank.

mstn commented 9 years ago

Sorry late reply. Do not worry if you are not able to map every line to gtfs.

5 is tricky. In general you cannot say where a circle starts or ends, because it is a loop! In order to follow the mapping "standard" of OSM (ie lines split into two directions), we picked two bold stops as endpoints. NB: bold stops in Miccolis' semantics mean:

  1. endpoint OR
  2. a place where arrival_time != departure_time OR
  3. a point of interest

If there are more than two endpoints, we should follow OSM. In OSM Line 5 is made of two pieces, from Piazza Matteotti to cimitero and back http://www.openstreetmap.org/relation/3776274 http://www.openstreetmap.org/relation/3776273

We have several trips:

In practice, you should care only about "regular" routes (Piazza Matteotti -> cimitero and back). We will add exceptions later.

It sounds complicated and it is! In my opinion those timetables are the results of ages of ad hoc changes. A sort of geocentric model. :)

mstn commented 9 years ago

Same rule for Linea 6A and Linea 6B (from OSM)

For example, via Messapi is the last stop. It is the last entry of the first column and the first entry of the second column, but with different pole numbers! Different pole numbers are not abstract metaphysical concepts, but actual poles. This means that in via Messapi the same bus stop has at least 3 poles, I guess. When there are double pole numbers (eg 21-22), it means that a bus stop has a brand new pole, but they "forgot" to remove the old one.

mstn commented 9 years ago

Linea 12 is Piazza Matteotti-Cimitero and back http://www.openstreetmap.org/relation/3776269

Note that Cimitero is bold.

albertocottica commented 9 years ago

Understood and fixed.

Now I have a problem with line 14, which follows two different routes on working days and holidays. The main stop is the same (Piazza Matteotti), but on working days the "waypoint" in bold is indicated as Via Arno; on non-working days it is Timmari. Provisionally, I am treating them as two completely separate lines.

albertocottica commented 9 years ago

Scrap that. On most runs, 14 on holidays runs as on weekdays, but there are exceptions. Man, these timetables suck!

It is 8:52. I have decent data as far as line 14, working days. I have to stop now, but will do my best to finish lines 14 and 15 by the end of the afternoon.

albertocottica commented 9 years ago

Quite difficult: 14 is done except on holidays – I added a separate line for corsa serale on weekdays 15 is too complicated, there are different runs having different ends. For lines 14 and 15 there are not always endpoints marked on bold. I have to stop now...

elf-pavlik commented 9 years ago

@albertocottica thanks for helping out with all those crazy time tables :exclamation: I think we should send some feedback to Miccolis and I hope that you can contribute to it. We arrange few days long sprint with @mstn in Trento Dec 6-12 and should get back to you about it then...

albertocottica commented 9 years ago

Absolutely, with pleasure!

On Fri, Nov 28, 2014 at 12:33 AM, ☮ elf Pavlik ☮ notifications@github.com wrote:

@albertocottica https://github.com/albertocottica thanks for helping out with all those crazy time tables [image: :exclamation:] I think we should send some feedback to Miccolis and I hope that you can contribute to it. We arrange few days long sprint with @mstn https://github.com/mstn in Trento Dec 6-12 and should get back to you about it then...

— Reply to this email directly or view it on GitHub https://github.com/unmonastery/GtfsMaker/issues/8#issuecomment-64837748.

elf-pavlik commented 9 years ago

@albertocottica let's maybe take an hour or two during 31C3 to wrap up on this one and most likely send some feedback to miccolis about their data and ask for further clarifications ?

albertocottica commented 9 years ago

With pleasure! I am even hoping to come to the sprint. If that happens, I volunteer to do the humblest tasks again – not that I can do much more!

elf-pavlik commented 9 years ago

@albertocottica @mstn can you take another look at line 14 and 15?

albertocottica commented 9 years ago

We discussed this with @mstn. The decision was that he needs to think about an improved procedure for me to input the many exceptions to the rules: first and last trips, schoolday variations etc. They might go directly in the GTFS file. Once he has decided, I can enrich the data files he chooses to work with.

elf-pavlik commented 9 years ago

Please continue in Matera specific repo: https://github.com/unmonastery/matera-gtfs/issues/7