Network constrained movement #307
Replies: 2 comments
-
Thank you for all these thoughts.
In general, 6-9 (and maybe 10) raise the larger issue of if/how we want to support network-constrained movement data. It opens up a whole new can of worms. The first step/hightest priority would probably be a proof of concept of a map-matcher/router integration to clarify the implementation details, e.g. assuming an irregularly sampled GPS track, does the map-matcher just snap the locations the the road graph and return snapped points or does it return the route along the road networks as lines? If lines, how should the timestamps be handled? I suggest to break this thread up by topic to make it easier to keep track: 1-3 on output unit handling; 4-5 on Pandas pyarrow implications; 6-9 on network-constrained movement; and 10 for faster stop detection. |
Beta Was this translation helpful? Give feedback.
-
Have split off 1-3 as an Issue #308, 4-5 as a separate discussion #309, and added discussion of 10 to the existing #296; so renaming this for discussion on network-constrained movement. As a first step, I'll look into map-matching and routing libraries to see if any would work OK. |
Beta Was this translation helpful? Give feedback.
-
get_length()
anddistance()
. Do any of these need the option to return measurements in different units?if name in DISTANCE_COL_NAME
. Probably doesn't need to be changed, but just thought I'd make the suggestion.traj.get_distance(units="km")
and the column name could return as "distance (km)". Easy to read the CRS units to have a default if someone just doestraj.get_distance()
.detector = mpd.TrajectoryStopDetector(my_traj, gdf)
and thendetector.get_stop_points()
ordetector.get_stop_time_ranges
would by default return only the points or time ranges at the locations listed in gdf.Would welcome any thoughts on these ideas. The last 5 come from the public transport data I'm currently working on, I am writing something that's specific to that because there's some things I want that wouldn't fit in mpd such as service area analysis etc. and bits that are specific to Ireland. But I thought these ideas above might be useful more broadly in mpd. If you don't think they'd fit, I can just write them for my project of course, but if you think they would be useful to have in mpd, there's no sense in me writing them just in my project code, so I'd be happy to contribute them and take suggestions.
Beta Was this translation helpful? Give feedback.
All reactions