Train arrival time algorithm

[jrootham]

Start by selecting all the samples for a train ordered by timestamp.
Merge all samples for a train in the same sample period to get a good estimate of the location.
Use those location estimates to generate arrival time estimates.

[radumas]

Hi @jrootham,

You can have at look at #13 for previous work on this. As well as doc/filtering_observed_arrivals.ipynb

[radumas]

I realize this isn't really written out in plain English in the notebook. The procedure I tried was:

1. Filter observations where the train is "At Station" or is <1 minute from arriving.
2. Order those observations by expected arrival time (create_date + timint)
3. If the traindirection is the same as the previous observation for this train, then it's the same trip. If if changes, then it's a different trip.
4. For each trip, the minimum estimated arrival is the arrival time, and the maximum estimated arrival is the "departure time"

Results.... could be improved upon.

[jrootham]

Hi Raphael I think we should talk about this tonight. Jim

…

On Mon, Oct 8, 2018 at 11:38 PM Raphael Dumas ***@***.***> wrote: I realize this isn't really written out in plain English in the notebook. The procedure I tried was: 1. Filter observations where the train is "At Station" or is <1 minute from arriving. 2. Order those observations by expected arrival time (create_date + timint) 3. If the traindirection is the same as the previous observation for this train, then it's the same trip. If if changes, then it's a different trip. 4. For each trip, the minimum estimated arrival is the arrival time, and the maximum estimated arrival is the "departure time" Results.... could be improved upon. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#40 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA5hfVrsuIWazaVq0u-HrIvSxiOuf8Rmks5ujBoWgaJpZM4XHXCk> .