What models and data is used for traffic predictions?

Question

I guess every country will have its own way of predicting traffic conditions. However, since global sites like Google and Tomtom, have live traffic services to get an impression of the current traffic situation, I suspect that uniform data and prediction models are used and available. What data is used, is it archived and how is it modeled? Maybe more important is it publicly available?

I would like to be able to predict the average traffic conditions at any given time on any given route. The only prediction I could find was from the Paris region, but that is only a visualization and not the actual data.

Answer 1

In this MIT project you can find a model about traffic prediction. Also, they mention that they took the data from Minnesota Department of Transportation. So, my guess is that if you want the data from a specific city, you have to contact with the local department.

Answer 2

You are looking at a multi-step process:

1. Identify data sources that provide an average annual daily traffic, a state department of transportation may provide this.

2. Identify sources for diurnal factors, in Florida, you could find a telemetered traffic monitoring site that has counts on an hourly basis similar to what the Paris region image shows. You can obtain samples of this data in Florida by going to the Florida Traffic Online site and looking for Red Dots, TTMS locations, and selecting continuous hourly counts.

3. You can then segregate the data by date, day of week, etc.

4. Now you can start aggregating volumes on a corridor or street network for periods of the year and days, balancing volumes accordingly.