Urban planners have a difficult task–to make recommendations on everything from zoning laws to highway routes to public transit expansion based not just on a city’s current patterns, but on the patterns that will emerge in the next ten, twenty, or thirty years. Imagine basing front-end web design or user interaction choices on the projected needs and assumptions of users who aren’t born yet.
To make life more complicated, obtaining data on human behavior in an urban environment isn’t as easy as, say, tracking a user’s path through a website. So when members of Finland’s urban planning community want to build pedestrian sidewalks or walkways through green spaces, they get creative. The goal is typically to minimize foot traffic outside the pedestrian walkways, preventing joggers and baby strollers from wearing down the grass.
Using Available Data
In order to build walkways that Finns actually walk on, planners had to find a way to observe behavior in the absence of existing visual cues. Existing walkways, as well as the worn-in dirt paths where frequent foot traffic had killed off the grass, were a form of data bias–they signal certain behavior to pedestrians who might otherwise have taken more direct routes or skirted certain areas of the park altogether. So the first step to data-driven sidewalk planning is finding a way to minimize or eliminate these sources of bias.
As it turns out, bias elimination in urban planning is only a snowfall away. A few inches of snowfall make it nearly impossible for a casual observer to distinguish between grass, sidewalk, and high-traffic paths worn into the dirt. And, even more conveniently, the snow makes it fairly easy to see the different paths pedestrians took across the park when “suggested” routes were absent. Convergence points appeared along the most-desirable routes.
To see desire paths in action, you don’t need to book a flight to Scandinavia. New York’s Central Park is home to a network of paths not included in original landscaping plans. Instead, planners waited for several months after the park opened and let high-traffic pedestrian routes appear organically. This way, they could bet that they were paving over places where New Yorkers would walk. Those not in a New York State of Mind can peruse the 1959 Chicago Area Transportation Study, which looks at public and private transit paths into the city’s bustling business district. Patterns differ between rapid transit riders, bus commuters, and car owners who choose to battle rush-hour traffic, but desire lines originating from a fairly wide swath of urban and suburban residential areas generally terminate within a single district.
Right, but what about the web?
In site-design terms, think of visitors as Chicago-area commuters. The odds are good that most of them come to your website with one of a handful of goals in mind. So to think like an urban planner, worry less about where, specifically, the traffic is coming from, and pay more attention to the places where users are wearing down the grass. Do they click through the navigation repeatedly before finding the page they want? Do they accidentally click on inactive page elements? As urban planners have come to realize, expertise is a poor proxy for the wisdom of crowds. If you can gather user logs (fairly simple), organize them, and analyze them (trickier, but still totally possible) you can let your users teach you what works best. It’s crowdsourced UX design, made possible by big data analytics.