91ΧΤΕΔ

Skip to content

UW NEWS

Engineering | News releases | Research | Technology

April 28, 2026

BikeButler map creates personalized routes for riders based on preferences like speed limits and road conditions

UW News

The interface of a bike-mapping app.
BikeButler is a demo web app that lets users find personalized bike routes in Seattle. Cyclists plug in their destination and origin β€” just like in other mapping apps β€” and can then toggle sliders for eight attributes to create personalized route options. Above is the interface. The images on the right show different segments of the route.

Even though he wanted to bike commute from his Capitol Hill home to the 91ΧΤΕΔ, Jared Hwang often took transit because he struggled to find a good bike route. Apps like Google Maps and Strava might suggest hilly, busy streets simply because they have bike lanes. He even headed to Reddit to crowdsource ideas.Β 

β€œI was like, surely, this cannot be the best way to do things,” said , a UW doctoral student in the Paul G. Allen School of Computer Science & Engineering. β€œThis data is out there. We know where bike lanes are, what the roads are like, what the speed limits are. We should be able to easily access all this information at once.”

So Hwang and a team of UW 91ΧΤΕΔers built , a demo web app that lets users find personalized bike routes in Seattle. Cyclists plug in their origin and destination β€” just like in other mapping apps β€” and can then create personalized routes by adjusting eight sliders.Β Β 

For instance, a cyclist can move a slider between β€œlow speed limits” to β€œhigh speed limits” or between β€œlots of greenery” to β€œno greenery.” The app generates route options based on those preferences. Users can then flip through images from segments of the routes and weigh the pros and cons of taking different streets. Notes on each segment tell users how it aligns with their preferences β€” for example, a three-block stretch might have low speed limits and good roads but no bike lanes.Β 

The team April 17 at the Association for Computing Machinery Conference on Human Factors in Computing Systems in Barcelona.Β 

Related

Researchers initially worked with four participants to understand how cyclists tend to plan their routes. Based on that, they built a prototype of BikeButler. For the basic street layout and other info, they pulled data from OpenStreetMap and government data sets. But those didn’t have information on more subjective qualities.Β 

For those, 91ΧΤΕΔers turned to Google Street View. They used a visual language model, or VLM β€” a type of artificial intelligence β€” to analyze street images and rate subjective attributes like greenery and pavement quality. The team had the VLM rate the level of greenery on streets and then compared this with two 91ΧΤΕΔers’ ratings. The humans agreed with each other about as much as they agreed with the VLM β€” about 60% of the time. Future 91ΧΤΕΔ might try to gather individual users’ greenery preferences to offset this discrepancy.Β 

Once they’d mapped most of Seattle, the team tested the prototype with 16 participants.Β 

β€œOverall the response was really positive,” Hwang said. β€œWe found that people do, in fact, have contextual preferences. A cyclist riding for fun on a Saturday might want a safer, greener route compared with their fast work commute. People intuitively know this, but it hadn’t been established through 91ΧΤΕΔ.” 

Researchers say future work might integrate feedback from the user study, such as the ability to drag routes to change them slightly and an option to take fewer turns. The team is currently studying how to quantify cyclists’ preferences around intersections and turns.

The 91ΧΤΕΔers note that the quality of BikeButler’s recommendations is constrained by the recency and accuracy of the data it uses. For instance, a new bike lane might not yet appear on a map, or it could appear in OpenStreetMap but not Google Street View. Also, since the team planned this as a proof of concept, BikeButler is limited to Seattle, though it could be expanded to other areas.Β 

β€œI’m a lifelong biker and bike commuter,” said senior author , a UW professor in the Allen School. β€œWhat excites me most about Jared’s work is how it points to a future where we receive route choices individualized to our preferences. So whether I’m biking with my two young children, or riding for groceries, I can find a route for that context.”

Co-authors include , a student at Issaquah High School and intern in the Allen School; , a UW doctoral student in urban design and planning; and , a UW student in the Allen School. This study was supported by the National Science Foundation.

For more information, contact Hwang at jaredhwa@cs.washington.edu.