How AI mapmakers are upending our understanding of the world

How AI mapmakers are upending our understanding of the world

Ecopia was the first to map every building in the United States. Now governments are turning to this Toronto company to help prepare their cities for climate change.


When the image first appears, it is as puzzling as it is striking. Geometric shapes in bright pinks and greens jostle for attention with bold lines that carve up the frame. The forms cluster together, but there’s no focal point. It could almost be modern art — perhaps a Mondrian, if he were over-caffeinated around neon paint.

But this isn’t cubism, it’s cartography.

“We’re over Italy,” says Brandon Palin with the authority of a knowledgeable tour guide. “You can tell by the structure of the buildings and the way the roads are.” We’re looking at an aerial view of a small town displayed on an enormous television screen in a downtown Toronto office. The neon-bright colours are from a digital overlay that obscures much of the photograph. It takes a moment to get accustomed to it, but then it becomes clear that every building, street, bridge, waterway, piece of vegetation and countless other features have been meticulously highlighted.

According to Palin, the somewhat haphazard arrangement of the houses is one of the clues that we’re in southern Europe. And he should know. He’s a senior director at Ecopia AI, a Toronto tech company that’s putting Canada on the map for, well … maps. Spun out of one of the co-founder’s research at the University of Waterloo in 2013, the company specializes in using artificial intelligence to create hyper-accurate digital charts from high-resolution satellite, aerial and street-view photography. Its technology is now deployed in more than 100 countries and its applications are endless. It is being used by urban planners in Italy, 911 dispatchers in Florida, and last year the company landed a multi-million dollar contract with the American government to help it prepare for climate change.

With the touch of a button, Palin transports us to a Seattle suburb to examine how much shade the tree canopy provides. Then we head over to Jacksonville, where the previously top-down map becomes three dimensional. This view is from the next generation of Ecopia’s technology, and the detail is impressive. The exterior shapes of the houses are rendered with such accuracy you can pick out the twee dormer roof windows beloved by suburban developers. The software can even determine what material the roof is made out of and assess its condition.

The capabilities of this technology are vast. It’s the world captured, digitized and ready to be explored.

 

Navigating the future

Climate change is intensifying the need for highly accurate maps. As sea levels rise and natural disasters multiply, governments, city planners and insurers need to predict the impact they might have as well as which solutions could help lessen the blow. Google Maps, the most high-profile player in the U.S.$20-billion digital mapping industry, is fine for getting from A to B and avoiding traffic on the way. But it can’t easily tell you which roofs may be suitable for solar panels or whether a patch of empty land is bare earth, vegetation or paved over. Those can be vital details if you’re trying to plan more clean energy infrastructure or determine where rainwater will end up in a heavy storm.

Jon Lipinski, Ecopia’s co-founder and president, says the company is on a mission to build a virtual replica — or digital twin, in tech-speak — of the entire planet that “reflects every detail of the real world and changes as they happen.”

That last part is crucial. A warming climate and fast-growing urban areas are accelerating changes in the environment. Palin points to communities in Florida that are being flooded as mangroves, which once protected them by absorbing winds and storm surges, are cut down for development. “The infrastructure put in place was designed back when those mangroves were still there,” he says. These days, mapmakers and planners have to run to stand still.

When high-resolution photography became widely available in the 1990s, extracting map information was a laborious task. Cartographers would trace the outlines of structures manually or methodically check and correct the work of some fairly hit-or-miss computer systems. The rise of artificial intelligence and growth in computing power over the past decade has enabled mapping companies to automate this process, crunching work that used to take weeks or months into days. Lipinski says the company’s secret sauce is its algorithms that enable it to extract enormous amounts of information quickly and accurately while also keeping file sizes down to levels that won’t melt its users’ laptops. That speed also means it can update its maps rapidly when new imagery becomes available.

 

 

Laying down markers

Ecopia’s technology has enabled it to notch up some notable achievements. It was the first company to chart every building in the United States and created the most comprehensive map of Sub-Saharan Africa. And it has also pushed its technology into mapping the vegetation like trees, forests and grasslands in unprecedented detail.

That work has attracted attention. The U.S. National Oceanic and Atmospheric Administration (NOAA) recently awarded the company a $10-million contract to provide maps to support its climate resilience planning, and Ecopia is now one of the go-to companies for state and municipal governments from Illinois to Los Angeles.

Here in Canada, the City of Peterborough is using Ecopia’s data in its stormwater models, with the goal of preventing a repeat of the 2004 floods that caused tens of millions of dollars of damage. Traditionally, hydrologists had to rely on rough estimates of how much rain various surfaces could absorb. Ecopia’s technology allows planners to more precisely model water flows by accurately identifying non-porous surfaces like roads and roofs and determining whether or not they’re connected to the city’s sewer system. The federal government has also contracted Ecopia to help plan broadband expansions as well as make detailed maps of Canada’s 100 largest cities to support net-zero efforts like improving transit.

The wide range of Ecopia’s activities reflect the fact that AI-generated maps are now the basis for vast amounts of planning work. “If you request a home insurance quote, make a call on a 5G network, or rely on a stormwater drainage system, all of those things have underpinnings in map technology,” says Lipinski.

In this way, AI cartographers carry on a tradition of mapmaking that shapes our world as much as represents it. Maps have always prioritized certain elements over others — the Romans located the Mediterranean at the centre of the world, while depictions like the Peter’s Projection have been produced to counter the gross over-exaggeration of northern hemisphere countries on most modern maps. By emphasizing detail and the need for constant updates, AI-generated charts push maps past their origins as static depictions of a slowly evolving world. They’re now dynamic tools to help us react to an increasingly complex and fast-changing planet.

“Maps are fundamental for understanding change,” says Palin. “Effectively, these data sets are the foundation for innovation.”

MaRS Momentum program works with high-growth Canadian companies to accelerate their path to hitting $100 million in revenue. Is your business Canada’s next anchor company? Find out more and apply to join the program.

Photos courtesy of Ecopia