TrafficMetrix®️ evolution: your need to know on the latest changes to Trafficmetrix®️

We’re always looking for ways to improve both our raw data and the service we offer to our customers. Our TrafficMetrix®️ product has recently been improved in a number of ways, with process automation playing a key role in making our datasets more robust, accurate, and reliable than ever. In this blog, Eleanor Michie, GIS Analyst at Kalibrate, answers questions about the changes and what they mean for customers.

What’s new with TrafficMetrix®️?

We’ve made improvements to the way we gather and process data, and added some new features to the product.

Firstly, we’re pleased to introduce our new Five Year Estimates. We’re now able to predict traffic patterns accurately up to 2025. These will be continuously updated with each quarterly release, to always offer estimates for the next five years.

We’ve also made some changes to internal data updates, automating a lot of the process to keep data constantly current.

We’ve shifted the timing of our quarterly releases, which will now take place in  January, April, July, and October. Previously, we were releasing in December, which wasn’t as useful for customers, given the Christmas lul.

Finally, we’ve refreshed our data for the US and Canada. All US States are up to date with the latest Department of Transport data as a minimum, with other data sources plotted where available. In Canada, every province and territory is up-to-date according to each provincial Ministry of Transport. This means we have unrivalled, accurate coverage of North America, built on actual recorded counts.

This is all in addition to the robust data sets we have available globally. In most countries outside of North America, published traffic counts are not available. In these cases, we manually collect traffic counts and utilize sophisticated modeling techniques to ensure accuracy.

Why is it important to use actual recorded counts?

We use actual recorded counts because it’s the only way to ensure total accuracy, but obviously it’s a big task to gather and process that amount of data. We’re the only provider on the market that does this, in fact. It’s certainly a large investment, but it’s something we pride ourselves on.

Other providers rely heavily on modelled counts derived from mobile data. Given not every vehicle is tracked, assumptions are made on the actual number of vehicles present on the road.

Additionally, where modelling is needed (5-year estimates and sites with no count updates), it is vital to calibrate and test models with the latest and highest quality counts. This is another important reason why we have invested time and skill in collecting and processing count data.

What’s new with how you’re processing the data?

Over the past six months, we’ve been automating as much data gathering and plotting as possible, so data is refreshed a lot more frequently. Previously, it would have been impossible to refresh all our data within a year, but now it’s refreshed almost as soon as it becomes available. The main benefit of this is the high level of accuracy it brings to our Current Year and Five Year Estimates.

What’s next?

We’re looking to develop a model that can apportion the total AADT across the two directions of the road. This will use machine learning combined with geospatial analysis to determine directional traffic where no observational data is currently available. This means that you don’t just know how many cars are using a road but also the direction in which they are travelling.

We’ll look forward to sharing these new developments as they become available! In the meantime, you can learn more about TrafficMetrix or request sample data for your area.