Laura Bliss is a staff writer at CityLab, covering transportation, infrastructure, and the environment. She also authors MapLab, a biweekly newsletter about maps that reveal and shape urban spaces (subscribe here). Her work has appeared in the New York Times, The Atlantic, Los Angeles, GOOD, L.A. Review of Books, and beyond.
A biweekly tour of the ever-expanding cartographic landscape.
Welcome to the latest edition of MapLab. Sign up to receive this newsletter in your inbox here.
Orient yourself: The map is an Ourosbouros
Google Maps, Waze, Uber, Yelp, and every other smartphone-based service you use to get around involves a fairly intricate map. That map may be interactive, offer navigation assistance, store your preferences, or connect to a network of users.
But fundamentally, it serves the purpose maps always have: to liaise between the reader and the world. Just as you must match a Thomas Guide or a subway map to your surroundings before making a move, digital maps still require people to glance up, compare screen to street or sidewalk, and make the next move.
Now, as the mapping industry booms alongside the development of autonomous cars, that model is changing. To brake, steer, and slow the car, autonomous vehicle software requires 3-D maps of the surrounding roads. What’s more, the machines are also generators of map data. Every time an AV is out for a drive, its sensors gather information about lane lines, curbs, and stop signs that are fed back into the master map.
Humans can incorporate inaccuracies and mismatches in the map as they chart their course. But now, “that ‘human as join’ model of maps is breaking down,” Young Hahn, the chief technology officer at Mapbox, told a keynote audience at Locate, a mapping and location data conference in San Francisco last week. Whereas “humans can look at the world and make the map work,” machines can’t, Hahn said. So that means that the data they gather had better be highly accurate, highly detailed, and at scale.
But what does it mean for a machine to draw and follow its own map of the world? Would the result be, eventually, a map that is impervious to inaccuracies and bias? Or is that pure “technochauvinism”? After all, the computer still has to match its map to the world, similar to the way that humans do. The artificial intelligence that “learns” the map could still be confused by minor changes, like a sparkly sticker on a stop sign. That could easily cause an accident or pedestrian injury—more easily, in that instance, than it would with a human driver.
In that case, what’s the purpose of a map that reads itself? And what potential subjectivities could still work their way in? These aren’t questions that you don’t hear asked much at tech conferences, but they’re ones that I’m pursuing right now. What do you think, readers? Let me know.
Compass points: Blockchain for good
Speaking of tech: Last week on CityLab, Sarah Holder and Linda Poon reported on how mapping apps and blockchain are helping cities tackle the growing homelessness crisis.
Holder writes for MapLab:
Conceptualizing the magnitude of the homelessness crisis in America is hard; curing its root causes is harder. But mapping its population, at least, is getting easier. In Spokane and Houston, a GIS-based “Counting Us” app allows the city to link its count of the unhoused to the real-time geographic location where they’re found. Patterns can emerge—where do people cluster during the winter, or after storms?—and care can be targeted.
Austin’s homelessness innovation resembles a map less in the traditional sense, but it does offer a way for people to keep track of themselves in a world categorized by impermanence. A blockchain-based database (crazy, I know) could soon hold a digital copy of homeless residents’ identity documents—important for getting a job, housing, or medical care—even without a physical home base.
For further reading, check out every county in the U.S. that has an affordable housing crisis. (Hint: it’s all of them!)
And here’s my 2017 profile of a formerly unhoused Portland, Oregon, entrepreneur tackling the worst of the housing crisis with a PadMapper-esque app specifically for the homeless.
Find all the bottles: a delightful map of Chicago’s Prohibition-era “ganglands.” ♦ Visit Milwaukee for the bratwurst, the lakeshore sunsets, and the country’s largest cartography collection outside of the Library of Congress. ♦ Not everyone gets arrested when they commit homicide in America. And not in every neighborhood. The Washington Post gives disturbing stats a compelling treatment. ♦ Useful new term alert: the “friction of distance” keeps people from economic opportunity. ♦ Mapped: why the impact of climate change will hurt poor countries most.
Don’t hog MapLab all for yourself. Send this to a friend and sign up for this biweekly newsletter here.