Every 40 days, a million people join Strava. It’s a staggering number for a niche social network aimed at athletes–but then again, Strava came along at the right time. The company, which was founded in 2009, has grown up alongside a groundswell of new cyclists and walkers in communities around the U.S. By some counts, commuting by bike has grown by more than 60% over the past decade, with walkers also increasing by leaps and bounds. While Strava’s core users tend to be athletes, it actively courts commuters as well: About half of the activities recorded through its app are commutes, rather than workouts.
Four years ago, the company launched a visualization tool meant to engage its users. It was a heat map–a mapping interface of route data that let you explore the frequency of routes taken around the world.
Strava was surprised to find that it wasn’t just users who were interested in the heat map: It was cities and states. “[We] had a lot of transportation departments reaching out to us, saying, ‘Hey, we could actually use a deeper dive into the same data, so we can better lobby for new infrastructure. We need to prove behavior change after new infrastructure is built, and we really need to drill down to [rider] counts, and temporal detail of time of day and day of week,” remembers Strava’s Brian Devaney. Soon, they found themselves figuring out a way to make all that data useful not for its athletes, but for planners and policy-makers. “That really launched us into figuring out, okay, how do we aggregate and anonymize and provide it back in a format that’s helpful for city planners?” Devaney says.
That’s how Strava Metro–tagline, “commutes count”–was born. This small branch of the company, which comprises about nine full-time employees compared to Strava’s 140 employees, was created to help cities make use of its massive trove of data about the way cyclists and walkers navigate streets. Devaney is the sales and marketing lead for the burgeoning wing of the app, which has seen similarly exponential growth. Within a year of the heat map’s launch, Metro was working with about 30 city and state agencies and departments to sell its data.
Today, Metro works with more than 125 agencies and departments, which pay a subscription-style licensing fee to Strava that is calculated per user (80 cents per user annually, which amounts to about $20,000 for Oregon’s Department of Transportation, for example). That includes major cities like Seattle and entire statewide groups like Florida’s Department of Transportation, but also smaller and rural areas like Rapides Parish in central Louisiana, which used Metro data to get its first comprehensive look at how people in the region bike and walk, resulting in its first bicycle and pedestrian plan.
Today, the company released its next iteration of the heat map that led to the birth of Metro. The new map is a reflection of Strava’s growth, with 31 different activity types that include swimming as well as conventional walking and riding routes. According to Drew Robb, an Infrastructure and Data Engineer who joined Strava five years ago, the new visualization required mapping 1 billion activities, or 13 trillion data points–roughly 10 terabytes of data. Created alongside the Metro team, the map is a reflection of the way Strava has created a de facto census of walkers and riders around the world–and how, with the right data design, that census can be incredibly useful.
To understand why Metro has grown so quickly, it helps to know how planners typically collect data about walkers and cyclists. Which is to say manually, using a clicker-style counter while standing on a street corner or at the foot of a bridge. While most transportation departments have a fairly fine-grained understanding of how cars or buses operate on their streets, walkers and cyclists are a very difficult group on which to collect granular, network-scale data. When a transportation planner or advocate has to go up against other groups to make a case for investing in infrastructure that champions people over cars, the sheer lack of data is hugely problematic. “A lot of bike and pedestrian planners just didn’t have any data at all to equip them to go to the table with and lobby for better infrastructure,” Devaney says.
Strava didn’t set out to fill in those data gaps, but by the nature of its large user growth, it was doing so in an ad hoc way.
When Metro works with a new city or state, it will usually spend a few months consulting with them to create a case study on what its data can do, providing a few months of sample data and developing examples of the kinds of problems planners can solve with it. Once the agency or department is up and running, they also have access to a forum where clients across the country can discuss the ways they’re using the company’s data. Devaney describes the range of groups that pay for access to its data as sitting on a spectrum. At one end, there are groups that simply need to prove to their constituents that people are riding and walking in their neighborhoods at all, and that they should invest in a master plan for pedestrians and bikes. On the other, there are more advanced groups that are trying to understand details like traffic trends and crash data, and analyzing the way current infrastructure is supporting walkers and cyclists. “Over the next few years, we would like to see our data really empowering groups to climb up that curve,” he says.
While only about half of its activities are commutes as opposed to workouts, Strava is actively courting commuters with campaigns like the Bike to Work Day challenge and campaigns like “commutes count,” which highlights the fact that many of the athletes that use Strava to log their workouts also use it to record their commutes. “We want people to know that your seemingly mundane activity that you’re doing every day to get back and forth to work. . . actually has the potential to improve your quality of life in your city,” says Andrew Vontz, communications lead at Strava.
Of course, there are inherent biases in Strava’s data. Not everyone in every city uses the app, and though it is free, it still requires a smartphone to operate–excluding people who cannot afford one. Others have pointed out the fact that people who exercise on bike, for example, tend to be affluent. Devaney points to Seattle as an example of a city that has taken this into account; the city compares its manual counter data to Strava’s data to observe the correlation between the two types of data. With a strong R-squared value of .91, it can be relatively confident that it can extrapolate a fair estimate of rider and walker activity on its streets based on Strava’s data. Though Strava’s users are an imperfect reflection of a full city of walkers and bikers, Devaney points out that it is still far more detailed than the data many agencies had access to previously.
The kind of data that Strava collects about its users isn’t so different from the data that developers or tech companies hope to one day glean from people who work and live in so-called “smart cities,” where sensors and cameras collect big data about how people move in their cities. But rather than imposing forms of urban surveillance on citizens, Strava’s users record this data willingly, and on their own terms. They’re free to make their activities private, and the data Metro collects is carefully aggregated and anonymized. Strava itself maintains a robust guide to privacy functionality–a solid user-centered UX practice that many apps don’t offer–that clearly spells out how to opt-out of sharing data with Metro, too.
In that sense, it’s pioneering a form of voting with your feet (or wheels); a vision of the smart city where the value of big data is shared between private company, public agency, and individual user.