Strava’s global cycling dataset represents one of the largest collections of real-world cycling activity ever assembled—but it tells a profoundly incomplete story. A significant portion of Strava’s data originates from just a handful of affluent, well-connected neighborhoods, leaving entire regions and communities essentially invisible in what is often treated as “the definitive map” of where people actually ride. This geographic concentration means that cycling infrastructure decisions, research projects, and investment priorities are increasingly based on data that systematically reflects wealthy, predominantly white urban areas while obscuring the cycling patterns of lower-income neighborhoods, communities of color, and suburban and rural riders.
The implications ripple far beyond data analytics. When city planners use Strava heatmaps to decide where to build bike lanes, they’re often amplifying existing inequities rather than addressing them. When researchers study “cycling trends,” they’re often studying the trends of people affluent enough to own a smartphone, subscribe to a fitness app, and live in areas with cellular infrastructure. When advocacy groups point to Strava data to demand new infrastructure, they’re inadvertently speaking only for a subset of potential cyclists—and potentially ignoring communities where cycling is most critical for affordable transportation.
Table of Contents
- Why Does Strava’s Geographic Concentration Exist?
- The Real Consequences of Missing Cycling Data
- How Smartphone Ownership Reshapes the Data Picture
- What Should Planners and Researchers Actually Do?
- The Broader Pattern of Algorithm-Driven Urban Inequity
- Examples of Strava Data Distortion in Specific Cities
- The Practical Limitation: Strava as a Dataset, Not Ground Truth
Why Does Strava’s Geographic Concentration Exist?
Strava’s user base skews heavily toward specific demographics and geographies. The app requires a smartphone and either a cellular data plan or wifi connection to record rides (though recordings can be uploaded later offline). Monthly subscriptions and premium features cost money. In many cities, this alone filters out lower-income residents who rely on prepaid phones, have limited data plans, or cannot afford app subscriptions. Additionally, Strava’s global adoption is not uniform—it became established first in wealthy Western countries and in affluent urban neighborhoods where cycling culture, disposable income, and smartphone adoption were already high.
The app also benefits from self-reinforcing adoption patterns. Early Strava users created social networks and challenges that made the app more valuable for subsequent friends and riding partners. In neighborhoods and regions where no one used Strava, there was no reason to start. In contrast, cities like San Francisco, Portland, and parts of London developed Strava cultures where sharing rides, comparing speeds, and joining Strava-organized challenges became social norms. This network effect locked in data concentration in places that adopted early, while many neighborhoods and regions globally remained data deserts.
The Real Consequences of Missing Cycling Data
When Strava data forms the foundation of infrastructure decisions, it creates a self-perpetuating problem. City planners in progressive cities like San Francisco, Seattle, and Amsterdam analyze Strava heatmaps to identify high-traffic corridors and install bike lanes along routes where Strava users already ride. These investments make those areas more attractive to cyclists, which increases Strava adoption in those neighborhoods. Meanwhile, neighborhoods where Strava adoption never took off remain virtually invisible to the algorithm-driven planning process. A heavily used but unmapped cycling route in a lower-income neighborhood will never show up on a heatmap, so it never gets infrastructure investment, which means cyclists in that area have no reason to switch to Strava to record their rides.
This bias is not purely hypothetical. Research on neighborhood economic disparities shows that lower-income areas of major cities often have less car ownership and more cycling and transit dependency, yet receive proportionally fewer cycling infrastructure investments. Strava’s data concentration has the potential to reinforce this pattern by making cycling activity in lower-income areas statistically invisible. A city planning team that relies primarily on Strava data may conclude that cycling demand is concentrated in affluent, educated neighborhoods—not because cycling is actually concentrated there, but because data collection is. The warning here is direct: treat Strava heatmaps as data about Strava users, not as complete data about cycling. Using Strava data to plan infrastructure without accounting for this bias risks making cycling infrastructure decisions that primarily benefit people who can already afford smartphones and data plans—while leaving actual transportation cyclists and casual riders out of the conversation entirely.
How Smartphone Ownership Reshapes the Data Picture
Smartphone penetration globally is not uniform, and even in wealthy countries, smartphone capability varies. A person might own a phone but have a limited data plan, an older phone that drains battery quickly while running Strava, or simply no interest in technology-mediated social fitness. In many developing countries, smartphone ownership is concentrated in urban areas and among higher-income populations. This means Strava’s global dataset is essentially a map of affluent, tech-engaged populations in major cities—not a map of cycling itself.
Even within cities, this creates visible gaps. In the San Francisco Bay Area, for example, Strava adoption and usage rates are dramatically higher in wealthy neighborhoods like the Marina, Pacific Heights, and areas of Oakland adjacent to tech hubs. Meanwhile, neighborhoods like East Oakland and the southern Excelsior district, which have high cycling for transportation, show significantly lower Strava activity. The differential is not because fewer people ride in those neighborhoods; it’s because fewer people in those neighborhoods record their rides on Strava. cyclists commuting to work on a tight schedule, cyclists who cannot afford or choose not to use data on their phones, and cyclists who ride at night or in less visible places rarely show up in Strava’s data.
What Should Planners and Researchers Actually Do?
Cities and researchers that want a complete picture of cycling patterns need to diversify their data sources. Strava data is valuable and can identify certain patterns, but it should be combined with traditional surveys, intersection counts, GPS data from bike-share systems, community engagement, and careful ethnographic observation. Portland’s Bicycle Advisory Committee, for instance, conducts regular community outreach in neighborhoods across income levels and explicitly asks residents about cycling patterns that might not show up in any smartphone app.
This approach is slower and more labor-intensive than downloading a heatmap, but it produces better decisions. Community engagement and local knowledge matter especially in neighborhoods where Strava adoption is low. Asking residents directly about where they ride, what barriers prevent them from cycling, and what infrastructure would actually change their transportation patterns reveals needs that data-driven algorithms miss entirely. This tradeoff—between the speed and scale of algorithmic analysis versus the depth and accuracy of community input—is worth accepting, particularly when decisions affect people’s mobility and equity.
The Broader Pattern of Algorithm-Driven Urban Inequity
Strava’s data bias is one instance of a broader pattern where algorithmic decision-making in cities tends to amplify existing inequalities rather than correct them. Ride-hailing algorithms that concentrate service in wealthy neighborhoods, predictive policing systems that over-police communities of color, property valuation algorithms that underestimate home values in Black neighborhoods—all follow similar logic.
Data that appears objective and scientific often reflects the biases and choices embedded in data collection, who gets counted, and what gets measured. The danger is that Strava data, precisely because it is large, public, and presented visually as heatmaps, lends an appearance of objectivity and completeness that it does not deserve. A city council member looking at a Strava heatmap might think they are seeing “where cycling happens,” when they are actually seeing “where Strava users ride.” The difference matters enormously, and it matters most to people already excluded from digital platforms and data-driven decision-making.
Examples of Strava Data Distortion in Specific Cities
In Montreal, Strava heatmaps highlight cycling activity concentrated in neighborhoods like the Plateau and the Marais, both affluent and cycling-culture-forward areas. Yet interviews with cyclists and transportation researchers reveal that many commuting cyclists use practical routes through lower-income neighborhoods that barely register on Strava because adoption there is low. Similarly, in London, Strava data concentration in central and inner-west areas (particularly wealthier inner London boroughs) misses cycling activity in outer boroughs and neighborhoods where cycling for transportation is common but smartphone-based fitness tracking is not culturally normalized.
The Practical Limitation: Strava as a Dataset, Not Ground Truth
Cities and researchers that use Strava data need to explicitly state this limitation in their reports and decision-making documents. A bike lane proposal based partly on Strava data should note: “This analysis includes Strava data, which reflects recorded activity primarily from users with smartphones and Strava subscriptions in these specific neighborhoods, and may not represent all cycling activity in the area.” This transparency matters because it signals that the decision-makers are aware of the bias and that they’ve considered cycling patterns outside the dataset when making final decisions. It also opens the door to community pushback—residents can say, “Strava doesn’t count us, but we cycle here anyway,” and that becomes part of the official record.
- —
You Might Also Like
- The Urban Cycling Noise Crisis Explained in One Statistic: Cyclists Report 85 Decibel Average Traffic Noise Levels
- The Cycling Insurance Gap Crisis Explained in One Statistic: Only 12% of Regular Cyclists Have Cycling-Specific Coverage
- The Cycling Connectivity Crisis Explained in One Statistic: Average US Bike Network Has Gaps Every 0.8 Miles
