What happens to a community with no "digital shadow" on the web?
Maps have always had a way of bluntly illustrating power. Simply appearing on one can be enough to make a place or community matter. Meanwhile, absence from "the map" conveys something quite the opposite. Recall 19th century colonial surveys of Africa with the continent’s vast interior labeled as “unknown.” That one word on unmapped territory was simply another way of saying – in the eyes of the mapmaker – that the region was of little consequence. Whoever lived there didn't matter.
This old idea of paper maps as power brokers offers a good analogy for how we might think today about the increasingly complex maps of digital information on the physical world that exist in the "geoweb." This is where Wikipedia pages and online restaurant reviews and geocoded tweets live, all theoretically floating atop the actual cities and neighborhoods they describe.
"In many ways, it’s not even an analogy," says Mark Graham, a researcher at the Oxford Internet Institute. "We actually are talking about maps of some sort, in a way. Just because they’re maybe more ephemeral, or maybe more invisibly layered over our cities, it doesn’t mean they’re any less important or any less real."
On these maps of digital information, a familiar trend is emerging. Some places are covered much more densely with information than others (Manhattan compared to upstate New York, Europe compared to Africa). But that information density bears no direct relationship to the density of human populations. And the gap between these two metrics provides a new way of looking at old questions of inequality.
Every technological innovation today around a new smart-phone app or web platform improving quality of life in cities comes with a caveat. What about the people who can’t access those tools? What about the people on the other side of the digital divide who lack access to home computers, Internet connections, unlimited data plans? These are the people who go "unmapped" in the geoweb.
Researchers like Graham struggle to measure this effect, in part because our concept of a static map is disappearing. Today, online maps are dynamic: They appear differently depending on when you view them, or where you view them from, or whether or not you’re logged into gmail while you do it. It’s increasingly hard, Graham says, “to get the sort of God’s eye view that you traditionally have when looking at a map of what’s out there and what’s being both produced and represented."
Graham and University of Kentucky researcher Mathew Zook (among the academics behind the excellent Floating Sheep blog) have been trying to find ways to capture what’s out there – or, at least, Graham says, "what’s codified, indexed and out there."
There are at least three ways to think about all of this digital information about real-world places in the geoweb. What types of content are out there layered over a city (FourSquare check-ins at its restaurants, Wikipedia pages about its parks, geotagged tweets from its residents)? Where is that content coming from (who’s writing those Wikipedia pages, those tweets)? And who’s looking at it all? "Basically," Graham says, "how visible are the digital shadows of cities?"
Some of these questions are more easily studied than others, depending on the platform. Wikipedia statistics can tell you how many people have viewed a given page this month, or how many pages exist related to a given place (its history, politicians, music festivals, etc.). This map, produced by Graham, shows, for instance, the density of edits to Wikipedia coming from Middle East and North African countries:
On this uneven landscape (averaging two years of quarterly data from 2010-2011), Israel produced almost as many edits (215,333) as the rest of the region put together (254,089).
Graham and Zook have also tried to measure information density in the geoweb by looking at what Google indexes about a place (after all, this is Google’s mission statement: to "organize the world's information and make it universally accessible and useful"). The process is a bit meta. "We just ask Google Maps how much it knows about any particular place," Graham says, "and it returns that information." The researchers essentially set up a computer script that visits hundreds of thousands of points on the world’s surface, conducts a search for indexed content related to that point on Google, then stores the results in a database.
"By aggregating all of that data," Graham says, "we can see interesting patterns about which parts of the world people are looking at and which parts of the world people aren’t looking at it."
The results show, for instance, that the Tokyo metropolitan region is more densely layered with digital content than all of Africa. This is a map of the world sorted by the density of information indexed by countries on Google Maps:
That picture raises the old question posed by those 19th century maps of Africa: What happens to the people who aren't on the map, or who are barely represented at all? "What I worry is that what this will start to do is simply reinforce the divides and the differences between the haves and the have-nots, the cores and the peripheries," Graham says. “It’s most worrying for the places that are essentially off the map – or not in the database.”
Think, for example, of a sandwich shop in a Detroit neighborhood on the other side of the digital divide that has no website, no Yelp reviews, no little red balloon on Google Maps. How do people find it? Surely this form of invisibility is bad for business. Graham offers up another example from his own research in East Africa, where he has studied how the Internet has impacted the tourism industry. He interviewed a Kenyan tour operator about why he recently started offering trips to Rwanda.
"He told me he Googled it. He Googled ‘gorilla tours, Rwanda,'" Graham says. The tour operator was pleased that Google enabled him to learn about new places and, as a result, to offer them to his customers. But of course Google doesn’t really generate information; it directs people to it. And so, as Graham pressed further, it turned out that the man was ultimately deriving his information from Wikipedia.
"This is obviously just one example, it’s an anecdote, but I think in a very real way that was sort of shaping the flows of capital and people over quite large distances between Europe, where they were coming from, and Kenya, and then Rwanda,” Graham says, "simply because someone had written a Wikipedia article about that place and not another place."
Graham often circles back to that example when he’s thinking about what it might mean for a place to not be represented, either in peer-produced information on the Internet or in the algorithms that sort it all. This happens at the national level, with inequalities emerging between countries and regions. But it also happens within cities at the scale of neighborhoods. Zook has conducted other research illustrating the disparities in information layered over racially segregated parts of New Orleans. And we can see similar patterns – sometimes deceptive ones – emerged in the tweets from New York City during Hurricane Sandy. During the storm, the densest quantity of Sandy-related tweets emerged from Manhattan, relative to other boroughs of the city. But that doesn’t mean that Manhattan suffered the worst damage; rather, that it often produces the largest quantity of data. It’s easy to conflate the two, though, which is why maps often equate with power.
All of this information, Graham stresses, doesn’t exist in some kind of virtual world that's separate from the real one. The two are intimately intertwined: We use digital information to navigate and understand the physical world, and in turn our experiences of place impact how we then contribute to the information about them. When we use this information (by, for example, clicking on a restaurant on Google Maps), we are often simultaneously consumers and producers of it (that single click is another data point in Google's vast machine). All of this means that the geoweb may not just be reinforcing real-world inequalities. In many ways, it’s also enabling us to have dramatically difference experiences of the same places.
Consider this last example from Graham and Zook’s latest research: They conducted Google searches for the word "restaurant" in Tel Aviv, in three different languages. This is the picture that turns up from the English-language search:
And from a search for the word “restaurant” in Hebrew:
And, lastly, from a search for "restaurant" in Arabic:
"In a very real way this information is encouraging or perhaps even making people experience fundamentally different cities," Graham says. And language is just one of the most obvious filters through which your experience of a city of neighborhood might be influenced on the web (if you use "social search" that draws on the feedback of your friends, the resulting information the Internet serves up would also look different).
So what’s the solution to all of this? How do we avert a world where beneficial new digital tools perversely wind up reinforcing real-world inequality, obscuring some communities while portraying others in depth? Graham doesn’t know exactly what this might look like, but it might help, he suggests, if the platforms that we use to access information did a better job of telling us what they don’t know. Think, for instance, of all the Wikipedia pages begging for more information. What if the whole geoweb were populated with empty placeholders that announced "something belongs here about the Rwandan tourism industry, but no one has filled it in yet"?
"This isn’t a new challenge in cartography, to sort of represent the unknown," Graham says. "Mapmakers have been doing this for centuries. But we look to maps because we want to see what’s there, not what’s not there."
The above image, from Google Maps, shows Times Square in New York City.