EDITOR’S NOTE: This is the first in a four-part series of posts this week by statistician Mike Logsdon. Mike bikes across the Fremont Bridge every day, and put together this analysis of the bike counter data “mostly for the amusement of friends and coworkers.” I hope you enjoy it as much as I do.
The Seattle Department of Transportation installed a bicycle counter at the Fremont Bridge on October 2, 2012. The Fremont Bridge is the primary route for bicycle commuters who live in northwest Seattle and work downtown. I’m sure that some brave souls tangle with the gnarly and unsafe Ballard bridge sidewalk, and some adventurous souls cross the Hiram Chittenden Locks, but the Fremont Bridge is by far the most suitable bike crossing for commuters traveling from northwest Seattle to the urban center.
The city of Seattle has made the hourly bicycle crossing data public, and uploaded it to Seattle’s data webpage. I accessed the Fremont bridge bicycle counter data on May 20, 2014. Note that this data can also be visualized at SDOT through a Tableau interface, if you wish to poke around in it yourself.
This document discusses the Fremont Bridge bicycle crossing data. First, we present the data, then investigate the following:
- How much does rain dissuade ridership?
- Does the Cascade Bicycle Club’s “Bike Month” affect the rate of bicycle commuting?
- Are Fremont Bridge bicycle trips increasing?
Figure 1 shows the daily total bicycle trips across the Fremont Bridge. Weekday bicycle trips averaged roughly 2,000 per day in the winter, and 4,000 per day in the summer, with daily variation of about 1,000 riders on each side of the average. Weekend trips were on average about 1,400 fewer than weekday trips. Federal holidays saw crossing rates similar to weekends. If you bicycled across the Fremont Bridge on a sunny, July weekday, it probably looked like this. If you did the same on a rainy, January weekend, you probably enjoyed a moment of peace and solace.
Figure 2 shows the entirety of the data, broken down hourly. This basic, exploratory view of the data confirms what we know should be true. (This is always heartening when working with data, as the opposite is often true, and you waste time searching for an error in the source data, and error in the computer code, or an error in your conceptualization of the project.) Peak commute hours are 8am to 9am, and 5pm to 6pm. Those peak commute times saw anywhere between 200 and 800 riders per hour, depending on the time of year. The weekends saw fewer trips, clustered more towards midday. I removed data from a few days with suspicious looking spikes, which have been acknowledged as erroneous, but one spike remained – 11pm to midnight on July 4, 2013. It would appear as though several hundred people rode bicycles across the bridge after the fireworks display.
Next in the series, we will look at the effect of precipitation on ridership volumes.
11 responses to “A statistical analysis of biking on the Fremont Bridge, Part 1: Overview”
Semi-random question. Why is there no bike counter on the university bridge? It seems like there should be nearly as many bikers crossing there.
bike counter and live video cam!
then we watch those countless southbound bicyclists trying to merge left across three lanes of speeding traffic heading up capitol hill. (aka the worst part of my commute.)
This seems pretty derivative. Was a weak prior holding you back from doing a Baysean analysis? If you are going to use a frequentist approach, Sounds like the ideal dataset for some sweet ridge regression analysis
Why not go for the best of both worlds – The Bayesian Lasso?
Bike data + Ridge regression = Bridge Regression?
I wish there were a like button on this blog.
Interesting that the evening rush hour commute trips are heavier than the the morning rush commute trips. Are folks taking a different route in the morning? Lots of folding bikes that folks bring on transit and then ride home? Frequent post-work socializing on either side of Fremont that takes people from other bike routes through here?
Kimberly– I would assume the last of your 3 guesses– that some of those that occur during “commute hours” are not commute trips but are instead people riding to dinner or something that’s not related to where or whether they rode in the morning.
oh, and also recreation. I think lots of people are likely to go out for their workout/exercise ride after work, and that might take them across the bridge.
Right, I was thinking of recreation, but not in the sense you note below where folks who might not have commuted at all would still get out later in the day for a ride. Good point.
Thanks for this work! Have you uploaded these Tableau vizzes to Tableau Public? I’d love to be able to start with that you’ve got and play with these!