Mapping a Route to a Better Navigation App

What makes a good digital map? We look under the hood and compare two of the most popular apps.
On September 14, as Hurricane Florence unleashed its fury on the mid-Atlantic states, websites from The Weather Channel to The New York Times featured eye-catching maps tracking the storm’s trajectory, charting wind speeds, and pinpointing affected areas. For sheer utility, however, it was hard to beat an interactive map that identified every open shelter, road closure, and evacuation route with up-to-the-minute traffic information. This map was published by Google.

For anyone who relies on their smartphone to get around (most of us), it isn’t surprising that Google would have the most detailed and practical storm map. Ever since the company’s Street View vehicles were spotted on America’s roads in the mid-2000s, Google has dominated digital maps, offering the richest, most precise, and most current location information for everyone from motorists to cyclists and transit riders, not to mention countless third party apps like Uber and Zillow. As of April, Google Maps had more than 154 million users in the United States—nearly seven times as many as Apple, according to research firm Statista. (The user base of Google-owned Waze topped that of Apple Maps by more than two million.) Globally, about two billion people use Google Maps.

“It would take Apple a long time to catch up.”

For Apple, Google’s massive lead has been a perennial source of embarrassment and irritation. And in 2015, it finally prompted the company to rebuild its maps from the ground up. Since then, Apple has been sending out vehicles equipped with cameras and lasers to survey the streets and collect the data needed to produce detailed and accurate maps, both in the U.S. and abroad. This effort will allow Apple to reduce its dependence on hundreds of third-party data sources and control its own information. The company began rolling out the new maps in July. The ultimate goal, according to Apple’s SVP Eddie Cue, is to create “what we hope is going to be the best map app in the world.”

Most mapping experts greet such a pronouncement with skepticism. Apple Maps will certainly improve, but many doubt the company will ever truly match Google. “It would take Apple a long time to catch up,” says Michael Goodchild, one of the world’s foremost experts in digital cartography and an emeritus professor of geography at the University of California Santa Barbara. Goodchild ditched Apple Maps on his iPhone more than three years ago after the app left him stranded repeatedly while on vacation in Hawaii. “There were a lot of missing streets,” says Goodchild. “I did a comparison with Google, and it was obvious they were way ahead.”

The reason Google will likely remain the leader in the field: mapping remains a largely manual, time-consuming process. Over the centuries, mapmakers have replaced tools like the compass, quadrant, and level with the GPS, laser, and radar. Other technologies—including sensors, image recognition, and artificial intelligence—are helping mapmakers work much faster than in the past. But in many ways, an accurate map is still a street by street, house by house, park by park effort. “It’s all fieldwork,” Goodchild says. Fieldwork aided by technology.

Most of us don’t realize it, but maps—even those printed on paper—are among the most information-rich visual tools. A city map, for example, starts with streets and bridges, parks, waterways, and boundaries. You’re also likely to find subway stations and bus stops, courthouses and hospitals, museums, and business districts. Established graphical conventions—size, font, boldness, color, shading—quickly allow us to distinguish main arteries from secondary roadways, downtowns from peripheral areas, sports arenas from shopping malls.

A good digital map must do all this well, and it must do it again and again at different levels of granularity so users can seamlessly zoom in and out. It must also offer additional information about stores and landmarks—reviews and hours of operation, for example. It must constantly update itself to display traffic conditions, road closures, and new construction in an ever-changing world.

Google understood early on that a versatile digital map, one that it could control and customize, was key to expanding its search from the web to the physical world. In 2005, the company launched Maps as a desktop service, combining data from a variety of sources including third party mapmakers, satellite imagery, and government databases. Soon, its quest for ownership of mapping data got a massive boost.

Street View began as a fanciful project, based on the outlandish notion that Google might capture 3D images of every street in America—and eventually the world—to create an app that could virtually transport users anywhere. Google’s mapmakers quickly realized the images contained data that would be invaluable, and quietly launched a years-long effort called Ground Truth. Using image recognition software and various algorithms, they began scouring the Street View imagery to extract street numbers painted on curbs, road and traffic signs, lane restrictions and information on rights of way, the names of businesses, government buildings, and landmarks. A team of human editors verified the data was accurate. As it gradually added layer upon layer of data, down to the outlines of buildings and homes, Google cemented its position as the maker of the richest, most detailed digital maps in the world.

Apple approached maps from an entirely different vantage point: as a navigation aid that would make its iPhone more useful. Initially, when the iPhone launched in 2007, Apple seemed content to license Google Maps. But as the relationship between the two companies soured following Google’s push into smartphones with Android, Apple decided to build its own mapping app. The initial result, introduced in 2012, was disastrous. Streets were missing, and the app sometimes sent users to the wrong place. Since then, Apple has done a lot to fix basic navigation. It’s also added layers of information about businesses, transit, and more. But the app still mostly relies on maps licensed from more than one hundred outside sources. That means Apple can’t always control accuracy. And when the company wants to fix something—add a new development, update a road, route users around a temporary detour— it can’t always do so (at least not quickly).

“Google Maps had more than 154 million users in the United States—nearly seven times as many as Apple.”

To address these problems, Apple got serious about overhauling its maps about four years ago. Nearly a decade after Google started collecting data for Street View, Apple sent its own vans on the no-longer-crazy quest to capture images of the world’s roads. Topped with sensors, lasers, cameras, and GPS, the company began creating intricately detailed renderings of streets and roads. And like Google’s Ground Truth effort, Apple began extracting addresses, business data, road signs, and more to build its own maps. The vans aren’t the only source of data (both companies use satellite imagery from third-party providers of geographical information), but they’re the cornerstone of Apple’s mapmaking efforts.

The results are promising. Addresses are becoming more precise, no longer loosely pegged to the location of a building, but instead to a specific entryway. Navigation for pedestrians is better, too, after Apple mapped sidewalks, walkways, and paths along parks. Parking lots are rendered in detail, so you can optimize your spot based on where you’re heading and find your car when you return. Where Apple used to display green patches for parks or recreational areas, you may now see a sports complex with baseball diamond, a soccer field, or a golf course complete with sand traps and blue reservoirs.

Navigation on Apple Maps has improved as well; the app now offers lane guidance based on your destination. It matches directions on the app with what motorists see on the highway, be it a road name or exit number on an off ramp sign. Apple also maps indoor locations in malls, transit stations, and airports far more precisely. The company can make fast updates, too. In late September, when San Francisco closed its brand new Transbay terminal after inspectors discovered cracks in some structural beams, Apple alerted users and routed them around it.

Of course, as Apple worked overtime to create a quality, detailed map, Google did not sit idly. While Apple’s vans were making their first drive, Google’s Street View vehicles were mapping and remapping millions of miles of roads across the globe, including more than 99 percent of those in America. As Apple is busy identifying a park’s walking paths, Google is pinpointing play structures and barbecue pits. Meanwhile, its algorithms powered by machine learning have gotten more sophisticated at extracting data from those images.

The company has also capitalized on the popularity of its maps to make them better, inviting users to be part of the process. More than 50 million Google Maps users around the world have contributed data through a service called Local Guides, entering corrections, information about businesses, details on bike paths and curb cuts, and countless other details. As a result, Google now can suggest a wheelchair-accessible path between two points and note when a business or restaurant has accessibility issues. A new feature, unveiled in September, makes it easier for a group of people to plan where to eat by allowing users to create a short-list of options and share it with others. Maps will also showcase popular hours and estimated wait times at businesses, or whether a cafe is known for loud music. Google’s huge number of user-contributors are an enduring and self-sustaining advantage; it lures more people to the app, which in turn leads to more user-generated refinements. “That’s a really hard advantage to overcome,” says James McQuivey, a vice president at Forrester Research.

“Making an accurate map is still a street by street, house by house, park by park effort.”

There’s a lot more happening behind the scenes in the making of digital maps, and few have studied the subtleties more than Justin O’Bierne, a cartographer who once worked at Apple. In a series of lengthy online essays, O’Bierne conducts side-by-side comparisons of the differences between Apple and Google. In the most recent piece, he dedicates some 60 pages to quantifying Google’s edge. In the process, he suggests that the company may have mapped virtually every structure (including garages and sheds) in cities and towns, often identifying even small details like bay windows. He explains how Google combines building outlines from satellite imagery with business data from Street View to automatically pinpoint commercial districts and other so-called areas of interest. “Google,” O’Bierne concludes, “has gathered so much data…it’s now crunching it together and creating features that Apple can’t make—surrounding Google Maps with a moat of time.”

It’s hard to know precisely how wide that moat is. But consider the following: Apple’s massive map overhaul—a years-long effort to catch up with Google—so far, has gone live only in the San Francisco Bay Area and parts of Northern California. The company plans to roll out additional areas in the United States over the next year, and its vans have only begun collecting street-level images in about a dozen countries. As Forrester Analyst McQuivey puts it, “No matter what Apple does, it will be a while before it shows up in Google’s rear-view mirror.”

Previous Article: