Historical GPS Data Use Cases

[rob_giggey]

Hello, I've recieved requests for the historical bus GPS data in the past and I'm working on the request.  I was asked for some of the use cases to help ensure we release the right data in the right way.

 

Does anybody have any specific things they would be looking to do with historical GPS data?

 

Thanks,

 

Rob

[swf...@gmail.com]

I've no specific plans, but generating stats for each stop, bus and time to get the distribution of how late you can expect the bus to be at a given time/place would allow a couple of things to be done:

1.  give OCTranspo feedback on changes they ought to make to the timetable: if a particular bus is always late they should change the schedule (hopefully they already do this).    For some routes stats on whether a particular service is likely to be running would also be useful.

2. This data, combined with the live data could be used in an app trying to find the best option for a particular journey.  It would also allow better estimates for the completion time for a journey -- this might be the most useful thing that can be extracted from the data.   Eg suppose you have to be somewhere at 5pm, an app could give the probability of each bus getting you there on time.   If live data is also available you can add that into the calculations.

John

[Kevin O'Donnell]

Similar to John's,

• Look "congestion" points in the network.
• For each trip, catalog the stop# at which trips tend to become late. IE:
• stops 1 to 20 are usually on time.
• from stops 21 to 27 trips tend to become late.
• from stops 28 to the end the trip is now "late" but completes the run without getting further behind.
• This identifies "congestion hot spots" in the city.
• Can factor for time-of-day and day-of-week.
• Visually, imagine a heat-map of the city, with "red" spots identifying "places where buses become late".
• The "spots" would actually be lines between two stops.

[rob_giggey]

This is great.  Keep them coming if there's more or iterations to these. 

[Jason]

I have a couple of things I would look to do with historical GPS data.

I'm working on a method of leveraging the data to recreate the missing shapes.txt file from the OC Transpo GTFS feed, or at least a reasonable facsimile of it. I am working on a method of this now using data from last September, but it's already stale. When I get it working, I'll need to keep it relatively up-to-date using the most recent data. I've published some early efforts, but I'm looking at some new ways of doing it now.

Second, I'd like to do some independent digging into OC Transpo's On Time Performance figures. The city publishes fairly high-level stats, but most people seem rather critical that the results reflect their experiences. I'd like to dig into it a little deeper to try to tease out the reliability distribution across space and time. This is my real reason for the shapes data above. :)

Finally, I'm hoping there's something we can use to help model the effects on/of the LRT, maybe the west-end extensions.

I'm sure others will come up with more practical, real-world uses. My own interests tend to the policy level. :)