It’s relatively easy to measure volumes and speeds. It’s harder to tell from that data if there is congestion at that location. Many companies provide city-wide congestion performance data for a hefty premium. I wanted to see if I could calculate the change in congestion by using public open data. The city of Austin publishes bluetooth travel time sensor data at a number of locations.
Travel time
To reduce this massive dataset I only looked at peak hour travel times, from the volume data I was able to assume a city-wide peak hour of 4:30-5:30PM. The dates for this analysis is March 13 to May 15. I pulled 2020 and 2019 weekday data for those date ranges. If we calculate a weighted average based on the number of samples, we get a 30% drop in peak-hour travel times.
Travel time is more useful when looking at individual segments. Check out the list on the bottom of this post for each segment’s change in travel time.
Travel time index
Travel time index is defined as the congested travel time divided by the free-flow travel time. A value close to 1 would mean there is no congestion, a value of 3 would mean it would take 3 times as long to complete that trip than if there was no congestion.
To estimate a free-flow travel time I pulled 2019 weekday 5:00-6:00 AM and 7:30-10:00PM travel times. I assumed that a majority of the data in these ranges would be close to free-flow conditions but also have a suitable amount of samples. Last year saw a travel time index of 1.61, post-lockdown this year was 1.04. This is a 94% drop in congestion as compared to a year ago.
If we group by corridor:
Check out the processed dataset here.
Travel time change by segment: