Hey;

I think you can find whatever datasets you are looking for at this website

https://datasetsearch.research.google.com/

and I think this is kind of the dataset that you're looking for

https://datasetsearch.research.google.com/search?query=%20traffic%20forecasting&docid=MvPepxWbMzMzrTCbAAAAAA%3D%3D

Yes, you can create for instance an agent based model (ABM) and generate synthetically a realistic traffic simulator. Make sure it has probabilistic choices included, to avoid some too artificial aterfacts.

I have worked with sets of particles interacting with each other with interaction fields in 1/r**n for n=1...7

That was for another purpose than yours.

Now you need to model your traffic, its heterogeneity, its multiple stakeholders.

Yes you can...see for example : https://www.researchgate.net/post/Is_there_a_way_of_generating_random_fake_numerical_data_that_would_fit_statistical_descriptions

This is similar to the Monte Carlo approach which uses probabilities to create a distribution from random data. It can be done in Microsoft Excel. Free and full copy articles explaining how to do this are linked here:

Article Prioritizing risks for the future using Monte Carlo simulati...

Article Playing risk roulette with Monte Carlo simulation using Micr...

I think real traffic can be very different from random traffic. Therefore, random data can only be used for debugging algorithms and programs. And real conclusions can only be made based on real data

Hey;

I think you can find whatever datasets you are looking for at this website

https://datasetsearch.research.google.com/

and I think this is kind of the dataset that you're looking for

https://datasetsearch.research.google.com/search?query=%20traffic%20forecasting&docid=MvPepxWbMzMzrTCbAAAAAA%3D%3D

Dear Muhammad Iqbal Hossain ,

Look the link, maybe helpful.

https://www.stata.com/support/faqs/statistics/random-samples/

Regards,

Shafagat

You can search some of the datasets here: https://data.world/datasets/traffic

For example, this one https://data.world/cityofaustin/cqdh-farx

Hey Muhammad Iqbal Hossain

Before proceeding ahead with any random dataset, I will suggest you go through Google's Dataset Search Engine at least once. It has got a massive 25 million datasets.

https://dev.to/anujgupta/google-s-25-million-datasets-a-perfect-gift-for-aspiring-data-scientists-3ekh

I hope this helps.

Yes, you can also use the Monte Carlo methods to create new random datasets. Make sure you do cross-validation of the model.