My talk at Berkman: Mapping a connected world

My colleagues rock.

When you give a talk at the Berkman Center these days, one of your rewards is a Berkman lunch apron. I like it so much that I’m now trying to schedule two more talks so I can get the Berkman spatula and spice rack.

(Next week’s lunch talk is about the Internet Safety Technical Task Force report authored by John Palfrey with a great deal of work by colleagues including danah boyd. Rumor has it the aprons they’ll be receiving are flameproof.)

Actually, the reason talks at Berkman are such fun is the opportunity to solicit feedback from a room full of extremely smart people. I gave a talk yesterday that touched on topics from cartography to phrenology, and I’ve been drowning in feedback and suggestions from friends for the past 24 hours. This is a very, very good thing, and a great reminder that the best time to give a talk on a topic is while you’re still figuring out what you think about it.

So, here’s the audio and video of the presentation, called either “Mapping Globalization” or “Mapping the Connected World”. (Globalization is an easier for folks to understand, while Connected World has less baggage.) I’m especially happy with the video as my excellent Berkman friends feature the slides primarily and video of me speaking only secondarily. Those slides are on Slideshare as well.

And here’s a quick outline and bibliography, either for those who don’t have an hour to hear me babble, or for folks who were hoping for more references during the talk.

I started by talking about a favorite map of mine, a “rebbelib”, which is a map of islands and ocean swells made by navigators in the Marshall Islands in the late 19th century. Made of sticks and shells, these maps are highly accurate in terms of island position – though scale is often distorted – but even more fascinating because they show phenomenon we commonly wouldn’t depict on maps. I love these maps because they show a critical piece of infrastructure – the ocean swells – which were critical to understand for navigation and for safe and efficient ocean travel.

Some references to maps from the Marshall Islands:
A detailed article from Visual Wikipedia
Academic references to articles on stick charts and to Captain Winkler’s 1901 article on rebbelibs
An excellent article on stick charts which features Marshall Islands stamps, which depict the charts.
An analysis of Winkler’s paper and his history of studying rebbelib
Another excellent overview of stick charts
A contemporary geographer looks at the charts and connects them to paddling songs, used to measure time

I’ve found (and been sent by many readers of this blog) a wealth of maps from the late 19th century that depict infrastructure: shipping routes, telegraph connections, railroads. It seems to me – though it’s hard for me to verify – that there’s a surge of interest in mapping infrastructure at the end of the 19th century, a moment in time where a great deal of infrastructure is being put into place, which helps enable a great wave of globalization. By some measures – particularly the mobility of individuals via migration – the world was significantly more globalized in 1900 than it is today.

It seems to me that these maps are a form of celebrating this infrastructure, and indirectly celebrating the connection it made possible.

The Mapping Globalization project at Princeton has some excellent early infrastructure maps.
Library of Congress’s collection of Railroad Maps.
One of my blog posts on mapping infrastructure.

While our age is characterized by international connectivity – made possible by infrastructures like fiberoptic cable, integrated electrical grids, air travel and container shipping – it’s surprisingly hard to find maps of infrastructure. This is especially odd because we’ve come to expect a huge amount of geographic data to be available via sites like Google Maps, and we tend to complain when these maps are censored (blurring Dick Cheney’s residence, for instance.)

Certain infrastructures have been mapped ad nauseum – there’s an explosion of internet mapping in the early 1990s, which peters out around 2004, perhaps because the internet is so complex that mapping becomes more of an aesthetic exercise than a geographic one.

Martin Dodge’s Atlas of Cyberspaces, 1997-2004
opte project’s internet maps, last rendered in January 2005
Telegeography’s undersea cable map

Maps of other types of infrastructure are available, but usually from commercial vendors, and they’re often not cheap. (I griped about the cost of Petroleum Economist’s maps on my blog recently, and they generously sent me a copy of the map I’d mentioned as well as their 2007 World Energy Atlas, a truly lust-worthy volume of maps.)

It is possible to reconstruct good maps of infrastructure if you’re willing to invest them time. John Young of Cryptome has created a very strange series of maps, called the Eyeball Series – some are overviews of critical infrastructure in the US, while others seek to reveal government secrets.

Spending too much time mapping infrastructure in the US can get you into real trouble. As a grad student at George Mason, Sean Gorman was mashing up a map of fiberoptic cables in the US with the locations of US businesses. The map revealed a large number of vulnerabilities, where it would be possible to cripple the communications infrastructure by attacking a small number of cables. US government officials took an interest in his work, making it very hard to distribute his thesis, and leading counterterrorism expert Richard Clarke to declare, “He should turn it in to his professor, get his grade — and then they both should burn it.”

If maps of infrastructure were celebrated in the late 19th century as icons of our newly global world, they’re feared today as maps of vulnerabilities. But while terror experts worry that mapping infrastructure can make us vulnerable, I worry that we rarely see these maps.

Generally, these maps get pulled out when something goes wrong. (See this blog post for a discussion of maps of infrastructure and infrastructure failure.) I worry that this distorts our understanding of infrastructure. To offer a (potentially inept, and very strange) analogy, the history of mapping human brain function incudes phrenology (mapping the visible geography of the brain by measuring lumps on the head) to mapping infrastructure when it broke (guessing at brain function based on massive trauma like the iron rod through the head of Phineas Gage.)

These days, we have a far more accurate way of analyzing brain function – we inject people with radio-tagged glucose or oxygen and map what parts of the brain light up on a PET scan. We’re mapping the flow of oxygen or glucose, but that flow allows us to intuit brain function.

To understand globalization and our connected age, we need to map flow, not just infrastructure. (Nearly all the examples in this paragraph are mentioned in this blog post.) A visualization of airplanes flying over the world shows us patterns that aren’t apparent from just looking at route maps – the dominance of domestic air travel over international; the thick connections between the east coast of the US and western Europe, an air corridor from Spain and Portugal to Brazil and Argentina, the almost complete absence of South-South air traffic. Maps of San Francisco made by tracking taxi cabs show non-obvious traffic patterns from neighborhoods to nearby hospitals, and render invisible neighborhoods where there’s virtually no taxi traffic.

Mapping flow is a form of surveillance. We’re able to do it when people cooperate (like the BBC project The Box, to monitor the flow of a shipping container around the world) or when we can blur the data (information on city traffic on Google Maps.) But it’s hard to map flow because people and things don’t stay still, and because they often – understandably – don’t like being watched.

Maps of flow aren’t the same thing as maps of intent – I may intend to go from Boston to Budapest, but if I’m flying Lufthansa, I’m going through Frankfurt. And as much as we may hope that infrastructure is built to match flow, that’s not always the case. Infrastructure precedes flow, and it rarely is able to adapt quickly to support new flows. There’s amazing economic opportunity in figuring out where flow exists outside of existing infrastructure and building infrastructure to accomodate it – the business case for small airlines in West Africa.

So what?

It’s my contention that most of us badly misunderstand what is and isn’t globalized in our connected world. French economist Daniel Cohen posits the phenomenon of “imaginary globalization”, where we see global stuff (a bottle of Fiji water, in my example) and assume we’ve got more connection with Fiji that we actually do. We don’t know much about Fiji, don’t know many Fijians, don’t get much news from Fiji, and Fijian ideas rarely shape our thinking.

When we don’t know about the infrastructure that connects us, we don’t know who we could be connected to and who we’re prevented from connecting to. When we don’t know what flows over this infrastructure, we can overestimate some kinds of connection (and underestimate others). To understand how the world really works, we need maps, not just of infrastructure, but of flow. We need maps not just of the internet and shipping lanes, but maps that help us understand who and what we pay attention to, how we get information, what we know and what we don’t know.

My objective isn’t to start a project to build an atlas of globalization – though if I did, it would be by linking to the maps others are making, rather than by becoming a cartographer. Instead, I’m interested in watching the ways people are building novel tools to allow citizens to build maps, opening up data and cartographic tools, and turning publicly available data into new maps. As we find ways to put these maps of infrastructure, flow and intention together, we’re going to learn something about the world and how we see it, much as a shell and stick map tells us volumes about the Marshall Islands a hundred years ago.