I am doing a research master's degree in transportation science. I would like to develop a model for one of my classes to detect the dependence between various bicycle counters.

The database I'm using is the city of Montréal's bike counters, which provides a CSV file with the location of all counters and another with aggregated countings every 15 minutes.

The format of my data is the following :

Date                  counter_01     counter_02     ...     counter_n
2022-01-01T00:00:00        0              1         ...          4
2022-12-31T23:45:00        3              2         ...          0

I want to be able to model the dependency that might exist between all of the counters, as well as the temporal correlation of the counter with its past recordings. The idea is to be able to predict or interpolate one or various counters if they are out of order for any reason.

I would like to incorporate the spatial dimension and, ideally, a network restriction (having the road network of the city of Montreal).

I have considered using BDLM or LSTM networks, but since I am relatively new to machine learning, I am not sure how I would be able to carry out my project with all the dimensions I want to put in it.

Any help or guidance will be highly appreciated!

  • 2
    $\begingroup$ Since you are "relatively new to machine learning" I would start with the basics. Choose a simple model and model one aspect of the problem, and then build from that. Or you might do some research if someone has done something similar. $\endgroup$ Jan 3 at 10:53


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy