"I AM TOLD there are some people who do not like maps. I find this hard to believe," says Eric Rodenbeck. The company he co-founded, Stamen Design, does interactive mapping projects that aggregate large amounts of hard-to-absorb data into visual images users can study by varying sliding controls and then draw their own conclusions. Instead, he says, of searching or scrolling through lists, users see patterns and anomalies in visuals right away. Stamen's goal is interested people, not experts. "By and large the field of visualisation and mapping has been stuck in this model of being tools for experts by experts." Recent projects track crimes in Oakland, CA and Mysociety's London travel times map.

Rodenback has always had a "map fetish". After architecture school, he went to work as the lead designer for Quokka Sports – " the only actual job I've ever had". A casualty of the dot-com bust, Quokka pioneered the use of the Web and new technologies to follow sports in new ways. Its most notable success was the Whitbread round-the-world sailing race. Watching boats come into and leave port isn't that fascinating, and therefore sailing is hard for TV to cover. Quokka, however, put GPS sensors and satellite uplinks on all the boats and contractually required competitors to transmit pictures, video, and 50 words of email a day to feed the service.

"We started putting it online 24/7 because we never knew when something interesting was going to happen and the sport has a global audience. So suddenly people could check in with their favourite team pretty much all the time. Mapping was a huge part in that." Rodenbeck left when the company "got too big" – and before it burned through its investors' capital and crashed. He went through a "dark period" in which he started a couple of companies that didn't work out well and kept his hand in with some mapping projects. And then he met Mike Migurski, Stamen's technical lead. "From there we started doing all kinds of interactive projects."

As Rodenbeck describes it, Stamen is the exact opposite of the old dot-coms. "We have an office at Mission and 16th, one of the worst blocks in San Francisco," he says. "It's very cheap to be here. We keep expenses really low so we can pick and choose projects." Stamen is particularly interested in working with cities. "Civic infrastructure is something that sounds boring," he says, " but people should have the ability to understand the place that they live in a way that makes sense to them. One of the things in the Oakland crimespotting project is that if you hear sirens on your street you should know why."

A typical project, he says, takes one to three months to build. "A month to figure out what to do, a month to build it, and a month to make it good." They let the data dictate the structure and try to come up with something that looks " interesting". Ahead of time, "We had no idea if there was a pattern in Oakland or Mysociety."

The hard problems Stamen encounters in these projects have less to do with data visualisation than with access to data: "All the GIS stuff out there is generally designed for GIS projessionals." Instead of formatting data so that only people who already know a lot about a city can understand it, Stamen's goal is to enable people to access data in the way they choose.

In the case of the Mysociety project, which maps London locations against the time it takes to travel to work, the goal was to avoid the more usual approach of requiring users to specify ahead of time. "We want to say, here is the whole landscape of possiblities, and give me a couple of dials and see the results as I make these changes." As a result, he says, "I've had people say that by looking at those maps they could find out in three minutes what had taken them three years of living there to find out." µ