How to predict Estimated Time for Arrival given only trajectory data and time?


I have data of latitude, longitude and timestamp. I am trying to build a graph based on pincodes (in India) (equivalent to zipcode). Based on this graph and trajectory data that I have, I want to predict ETA given source, destination and time for dispatch.

I am very new to this field, So I dont know how to handle and make structures using this data.

I have already gone through this question Learning time of arrival (ETA) from historical location data of vehicle but it is not very clear to me. Any kind of help/suggestions is appreciated.

Thanks in advance.

Edit : Pincodes are equivalent to zipcodes.

Graph : Nodes of the graph will be pincode boundaries and if these pincodes share boundary then there will be edge between them.

Trajectory : It is the data collected from GPS when vehicle is travelling on the road. Data is in format (Latitude, Longitude, TimeStamp, vehicleId)


Posted 2016-02-09T11:22:31.957

Reputation: 99

I think you should clarify what pincodes means, what graph and trajectory mean in this context. I'm not sure what the problem is – Sean Owen – 2016-02-09T19:16:40.473

Hi Sean, I have edited the question with details you asked for. I am new to the forum so kindly tell me if something more is expected in question details, so you and other users can help me out. Thanks. – user825828 – 2016-02-10T21:45:57.010

No answers