GOOGLE MAPS PIONEERS EXPLAIN WHAT’S NEXT, WITH PREDICTIVE MAPPING IN REAL TIME

by Ariel Bogle

Google Maps won. Look at your smartphone: it is likely to be one of the few apps you open daily, alongside Facebook and email. Now after remaining one of the most downloaded apps year after year, and convincing us it is indispensable, Google Maps wants to get even more personal.

Although other products may receive more attention, including Google Home or Waymo, the company’s self-driving car project, Maps is quietly remaking itself as a real-time engine of nudges and hints.

Consider the note in the app alerting you that a location may be shut on arrival, beta testing of a tool indicating whether there is likely to be parking at your destination, or even a feature that shades city blocks if Google thinks they are “areas of interest”.

It is suggestive mapping, although Google’s developers would probably prefer you called it “being helpful”.

Google Maps launched in 2005 after Google had acquired Sydney-based mapping start-up Where 2 Technologies in 2004 and made it the basis of its local engineering team. Of the original team, Andrew Foster is one of the few survivors. A group product manager, he has worked at Google for more than 10 years.

“Very early on when you went to Google Maps there was one or two countries and the rest of the world was water – it was pretty clear everything had to be filled in,” he says, suggesting the project vision hasn’t changed so much as expanded.

The intention was always to reflect the real world, but the level of detail has necessarily increased over time: “It’s not just there’s a coffee shop at this place and it’s open at seven in the morning. It’s ‘does this coffee shop have wheelchair access?’ ‘Free Wi-Fi?’ ‘Is this a great place for organic food?’.”

Although Google Maps has always claimed a degree of truthfulness about representing the world as it is, the map appearing on your screen is only a sliver of what its servers know.

Deep knowledge

Called “Ground Truth”, Google’s internal map is built from a mix of its own satellite, aerial and street view imagery, as well as an array of government data sources.

In many ways, choosing what and how much to show users is a matter of space. For Casey Whitelaw, a director of engineering on the Maps team, the challenge is finding the right ways to give people information at the right time without overcrowding the screen.

“I can think of lots of kinds of data that you don’t often see in Maps that we have underneath,” he said. “There’s too much information if we showed you every street number along a street on a map, or if we showed you every single shop. There just wouldn’t be room.”

Ground Truth is the portal to a much larger future. Its pinpoint accuracy could become the type of map necessary for autonomous and semi-autonomous vehicles – a project Google is heavily invested in.

If you are driving and want turn-by-turn navigation, for example, you’ll want to know exactly where to be on the road without extraneous information. “These days we’re getting to the point that we know which lane you should be in,” Whitelaw said. “We just didn’t need to know that before.”

But the immediacy and detail we now rely on could never have been achieved from a desktop computer. The smartphone allowed Google to know your exact location and build a hyper-specific reality around you.

Whitelaw suggested the smartphone brought extra demand from users, too. “Instead of it being a research tool, it became something you used in the moment,” he said. “It was, ‘where am I now? What’s around me now?’.”

Still Google’s attempt at “predictive truth” can’t be done without your participation. As the technology writer Alexis Madrigal points out, you’re the data point that makes the whole thing run.

Human intelligence

Take real-time traffic, for example. When you see a highway turn red up ahead on the map, indicating a traffic jam, that’s based on signals from users as well as some trend tools, according to Ramesh Nagarajan, a group product manager at Google Maps.

It can happen manually, too. Each photo you add of a restaurant, or each time you let Google know a shop has closed for good, you’re encoding their product with human intelligence it can use to serve someone else. Think of it like Amazon’s “customers who bought this item also bought” algorithm, but for the best way to walk a city block.

There’s little question that many of us decided long ago to see the labour of being data mined as a disquieting yet fair exchange. In some cases, it’s one we enthusiastically participate in. As Whitelaw put it, “whenever you give people a bit more information, they’re just voracious. They just come back to ask more and more.”

“You can tell me it’s usually busy on Tuesdays, but what about this Tuesday?” he adds. “Things like events and road closures and parades – whenever you give people a little bit more of this, they want more and more.”

Nagarajan says his team are especially focused on commuters. Although most people know how to get to work and won’t necessarily need step-by-step instructions from Maps, Google is still thinking of ways to boost its input.

“There are delays, the buses are offline today, there are some alternates to the schedule, there’s heavy traffic,” he says. “Those are the types of things we want to make sure we analyse and proactively assist the user [with].”

In February, the company announced a new Google Maps feature on Android that gives users a real-time ETA if they have their home and work addresses saved, as well as bus route and restaurant recommendations.

A criticism sometimes raised against Maps is its potential to homogenise experiences, by directing everyone the same way but Foster says he believes most people feel the service improves and guides, rather than defining choices.

“I think people become reliant on Maps in certain ways, but I think the positives outweigh that pretty strongly,” he says.

“If you look at it, over 11 million Australians are using Google Maps to get to work, so I think many people who work on Maps feel that sense of responsibility.”

After all, they’re part of the tiny, suggestible data points that make up Google Maps too.