Looping you in on travel time data

by Mike Allende,

The recent run of extra-challenging slow morning commutes – especially southbound out of Everett – has made travel times a big point of discussion. With several commutes topping the 100-minute mark – topped by a 140-minute time in late September – our travel times page has been getting a workout.

Travel times posted on our website let people know what their
commute looks like before they leave their home.
They update every five minutes.
There are a number of reasons why this is happening, from dark and wet conditions, to collisions and breakdowns in the wrong place at the wrong time, to simply a lot of people going to the same place at the same time every day. But that’s not what this blog is about.

Instead, I thought it might be interesting to take a look at just how we compute those travel times. We know people rely on them to plan their trips, and we use technology and data to make them as accurate as we can. It’s important for us to provide useful data to the public. So how does it work? Glad you asked.

Travel times listed on highway message boards give commuters
an idea of how good (or bad) their commute ahead is.
About every half-mile or so on highways in the area, we have loop detectors embedded in the pavement. These loops measure the speed of each vehicle that goes over it, and the amount of time that vehicle is on the loop. It then sends that data to junction boxes nearby which calculate the information and sends it to our website, giving commuters the amount of time it should take them to get from particular points. Our data is accurate about 95 percent of the time and updates on our website every five minutes.

But we don’t just rely on our loops. From time to time, we also have people drive the various routes at different times of the day to calculate travel times. We’ve found that the results are usually close to what our posted travel times are, which gives people a good idea of about how long it will take them.

Each month, we look at data from the previous three months to come up with the average travel time. If you watch our travel time page closely, you’ll notice that the average travel time changes throughout the day. That’s because we come up with averages based on time and day of the week, so the average time for the drive from Everett to Seattle at 8:15 a.m. on a Tuesday may be different than the average time for the same route at 7:35 a.m. on a Thursday.

Loop sensors embedded in the pavement of highways measure
the speed of each vehicle going over them, which are then
converted to travel times.
Something else to keep in mind is that it’s hard to account for poor weather when it comes to travel times. Rain and ice changes driving conditions in a big way and makes commutes much less predictable. A collision or stall blocking a lane can be exacerbated in bad weather and all of a sudden a free-flowing drive can become jammed within a few minutes, though that may not be reflected right away in our travel times. That’s why it’s important to check in with other sources, from our @wsdot_traffic Twitter account, to our Seattle traffic page and the media to get updates on road conditions. If you’re driving, of course, stick to radio traffic reports – please don’t check Twitter or our traffic page on your mobile device if you’re behind the wheel.