Skip to Content
Uncategorized

Killer Maps

Google, Microsoft, and Yahoo are racing to transform online maps into full-blown browsers, organizing information – and, of course, ads – according to geography. The likely winner? You.
October 1, 2005

Say you’re a fan of Paul Cezanne, and you’ve decided to fly to the Windy City to see his masterpieces at the Art Institute of Chicago. The only problem: you’ve never been to Chicago and have no idea where the Institute is or where to stay or eat.

No worries. You fire up Google Earth on your PC (it’s a free download at earth.google.com) and enter “Art Institute of Chicago” in the search box. Off to one side of the screen, a traditional list of search results appears, with the Art Institute at the top.

But the real action is in the central window, where Google Earth’s virtual camera zooms from an orbital view of North America down to a satellite image of downtown Chicago.

You click on the top listing in the search results, and a balloon pops up over the image, indicating the Institute’s building and giving its street address, a Web address, and a link to driving directions.

You plan to rent a car at O’Hare International Airport and drive to the museum, so you click on the “directions” link. Once you’ve entered your starting point – in this case, ORD, for O’Hare – the camera zooms out a bit and shows a colored line marking your route. Turn-by-turn instructions appear on the left.

You’d like to rehearse the drive, so you click the play button for an animated preview: the camera swoops down to a one-kilometer altitude and flies along the entire route, which turns out to be an easy 27-kilometer trip along I-90 and a few surface streets.

You still need a hotel, so you click on the “Lodging” check box. The map fills with options, and you spot a Crowne Plaza just around the block from the Art Institute. The hotel’s balloon leads you to Google Local, which displays the hotel’s location on a standard Google map and gives links to customer reviews, summaries from Frommer’s and other travel guides, and sites where you can book a room.

You reserve a room and go back to Google Earth. Now you can search for restaurants by clicking on the “Dining” box – or you can click on “Buildings” to get a 3-D view of Chicago’s cityscape. You notice that the Sears Tower is about eight blocks from your hotel; being an architecture buff, you add the 103rd-floor Skydeck to your sightseeing agenda.

You want to see what the area around the institute really looks like, so you leave Google Earth for a moment and visit geobloggers.com. There, you can use Google maps to search the popular photo-sharing site Flickr for shots that have been “geotagged,” or encoded with latitude and longitude information. On a Google map of downtown Chicago, you now see numerous pushpin-like place markers; as you click on each one, a balloon appears with a thumbnail image linking to Flickr.

Finally, you’re wondering how safe the area is at night. You leave Google Earth for a moment and visit chicagocrime.org, a free site that uses Google maps to show the locations of crimes reported to the Chicago Police Department. You zoom in on the intersection of Wabash and Madison, near your hotel. A handful of incidents show up on the map. You decide that caution is in order, but that there’s no reason to be paranoid. When you finally get to Chicago, you’ll be fully prepared – and you’ll be able to view Cezanne’s azure Mediterranean seascapes and luminous still lifes with a clear mind.

Read/Write Maps
There was a time, not very long ago, when the best way to prepare for a trip like this one was to order one of AAA’s custom TripTiks, map flip-books in which actual humans traced AAA’s recommended routes by hand with highlighter pens. You can still order a hard-copy TripTik, but since the mid-1990s it’s been faster to build it yourself on the association’s website or visit a site such as MapQuest.

And now consumers have access to advanced geographical visualization tools such as Google Maps, launched by the search giant in February, and Google Earth, released in June. With their combination of detailed aerial and satellite maps, high-powered graphics and animation, comprehensive local search functions, and hackability – it’s child’s play for programmers to display their own data atop Google maps – the new programs make paper maps and previous generations of online mapping tools seem primitive.

“I describe it as a browser for the earth,” John Hanke, general manager of Google’s Keyhole group, says of Google Earth. Keyhole, where Hanke was CEO until Google acquired the company last year, developed the software upon which Google Earth is based, mostly for customers in defense, engineering, and real-estate investing. Now that Keyhole is part of Google, the idea is to use geography as a fundamental structural principle for the entire Web. “The interesting part is not necessarily the core map but the information from the Web that’s now being organized geographically, so that you can get to it and understand it in its proper context,” says Hanke.

It’s such a potentially lucrative idea, in fact, that Microsoft has followed suit, introducing its own search-and-mapping service called MSN Virtual Earth. The service offers satellite photos, zooming and panning abilities, and interactive search listings resembling those of Google Earth, but it may actually reach a wider audience than Google’s product, since it runs inside a browser window rather than needing to be downloaded as a separate application. Yahoo, too, is in the game: last year it introduced maps that provide, say, the locations of all the coffee shops with Wi-Fi hot spots within a particular neighborhood.

Crucially, each company has released instructions for out-side programmers – called application programming interfaces (APIs) – that let them build online services that tap into the company’s own map programs. Developers are taking advantage of the new APIs to put out geospatial applications such as geobloggers.com and chicagocrime.org. The fact that these “mash-ups” are so easy to make is giving rise to a community of mapmakers and map users who are busy geotagging every piece of place-related information they can put their hands on. And the more information on the Web that’s tied to geographical coordinates, the better the results – and the better targeted the ads – that can be served up in response to location-driven searches.

The mapping revolution could, in short, change the way we think of the World Wide Web. We’ve long spoken of the Web as if it were a place – with “sites” that we “go to” – but as places go, it’s been a rather abstract, disembodied one. Now that’s changing. Geotagging means the Web is slowly being wedded with real space, enhancing physical places with information that can deepen our experiences of them and making computing into a more “continuous” part of our real lives (see “Social Machines,” August 2005).

For example, users of smart phones and wireless PDAs with location technologies such as Global Positioning System chips may soon be able to automatically retrieve stories, photos, videos, or historical accounts related to their current locations, along with ads and listings for nearby shopping, dining, entertainment, and business outlets.

And the information is already flowing both ways: users can upload their own texts, photographs, and other data to the Internet and pin them to specific latitudes and longitudes. “Historically, maps were a ‘read-only’ medium,” says Schuyler Erle, chief engineer at Locative Technologies and coauthor of Mapping Hacks. “Maps were only created by professional cartographers and professional GIS [geographic information systems] people. What has happened because of Moore’s Law is that people now have the computing power on their desktops to manage the vast amounts of data that are required for digital cartography. Maps are increasingly a ‘read-write’ medium. That changes how we interact with them and the impact they can have on our everyday lives.”

Many details of the emerging geospatial Web have yet to be worked out. No one knows which location-finding technologies are right for consumers or which will be endorsed by cellular carriers and device makers. Only a few of the U.S. cellular networks currently sell phones with GPS chips, and only one, Nextel, actually makes its phones’ GPS functions accessible to software developers. Outside North America, the mapping revolution may take longer, since some foreign governments maintain strict control over map data or charge exorbitantly for it.

But none of that is damping excitement in the community of Web developers and e-commerce managers. In June, more than 500 executives, programmers, and professionals, including some from traditional GIS companies such as San Diego’s Environmental Systems Research Institute (ESRI), converged on San Francisco for Where 2.0, a new conference organized by tech-book publisher O’Reilly Media. Participants spent two days admiring one another’s latest mapping creations and strategizing over how to convert geographic information delivered over the Web into actual transactions – from simply clicking on an advertisement to buying a house.

Trailblazing
How did such advanced mapping tools wind up in the hands of average Web users?

The short answer starts with the U.S. Department of Defense’s 1978 launch of the first satellites in the Global Positioning System. A GPS receiver determines how long the time signals broadcast by several GPS satellites have taken to reach it and, with a bit of spherical geometry, can then calculate its position to within a matter of meters. The original use of the system was to allow U.S. missile submarines to determine their positions within a few minutes of surfacing – information required by the guidance systems in the subs’ ICBMs if they were to make direct hits on enemy missile silos.

Beginning during the Reagan administration, civilians could also use GPS, but only in a degraded form, accurate to about 100 meters in any direction. On May 1, 2000, the intentional degrading of GPS signals, called Selective Availability, was turned off by order of President Clinton, instantly reducing the range of error in a civilian GPS fix to 10 meters or so. This sudden and enormous increase in the accuracy of GPS location-finding set the stage for all the online mapping innovation that has followed.

It spurred a broad group, including hikers, hackers, and urban planners, to take a deeper interest in Web-based maps, which were a natural way to publish the new geographic data they could collect and share using their GPS units. After all, a consumer-grade GPS receiver could now distinguish between one side of a street and the other, determine which storefront a user was walking past, or guide someone to a hidden “geocache” using only its published latitude and longitude (see “Roamin’ Holiday,” September 2005).

Unfortunately, when it came to making online maps, there weren’t a lot of options to choose from. Since its launch in 1996, one website – MapQuest – had dominated this niche. And while many Web developers wrote programs that copied MapQuest maps for redisplay in other contexts, they couldn’t program more sophisticated tricks, such as overlaying their own data on MapQuest maps.

“The first-generation Web services in the mapping space – ESRI, MapQuest, MapPoint – have had APIs for quite some time, but they weren’t hacker-friendly,” says Tim O’Reilly, CEO of O’Reilly Media and creator of the Where 2.0 conference. Eventually, MapQuest prohibited even the repurposing of its maps. This created a demand for reusable map data, a demand that would eventually be met by companies such as Google, Yahoo, and Microsoft.

Along the way, however, a few other things had to happen. First, computers needed enough processing speed and storage capacity to handle the multigigabyte data sets and complex mathematical transformations that displaying and manipulating digital maps require. As Locative Technologies’ Erle notes, Moore’s Law took care of that.

Second, the sharing-oriented mindset of the open-source-software community, along with an awareness of the possibilities of the Web, had to penetrate the walls of traditional GIS companies like ESRI. ESRI had long focused its products on industries such as financial services, urban and regional planning, and defense. Its emphasis, understandably, was on building accurate maps to convey critical data, not on tinkering with code or putting fun, interactive maps on the Web (see “Do Maps Have Morals?” June 2005).

But over the last several years, conversation within industry standards groups like the Open Geospatial Consortium (of which ESRI is a leading member) and the World Wide Web Consortium at MIT has led to agreement on basic standards for mapping-software APIs – and on additions to the Web’s central language, XML, that make it easy to tie Web documents to geographical locations. Embedding the XML tags <geo:lat>38.888</geo:lat> and <geo:long>-77.035</geo:long> in a Web document, for example, lets mapping or browsing software know that the document is about the Washington Monument.

Third, owners of large, valuable, proprietary databases on the Web needed some time to arrive at the idea that granting outside access to their databases might actually be good for business. Amazon was one of the first companies to put this idea into practice, releasing an API in 2003 that allows programmers to tap into its product database, pull out whatever information they want, and present it on their own websites in any format they choose, as long as any resulting purchases are directed back to Amazon (see “Amazon: Giving Away the Store,” January 2005).

The basic idea of Web services – that the software and databases powering e-retailing, online photo-sharing, and the like should be built according to standards allowing other parties to tap into them – was still radical even three years ago. Today, however, it’s the guiding principle of an increasing number of open-source developers and megacorporations – even Microsoft.

By early 2005, then, the hardware, the standards, and the collaboration models were in place for a burst of innovation in Web mapping applications. All that was needed was a starting gun. The gun fired on February 8 – the day Google Maps went online.

Map Mash-Ups
Even on the surface, it’s clear that Google Maps goes much further than older interactive map sites. The stunning satellite views, along with the ability to drag the map in any direction without having to wait for the page to refresh, are the most obvious advances. The shaded pop-up balloons pointing to the locations turned up in local searches – Google calls them “info windows” – are also a pleasing touch.

But it’s what’s under the hood that really excites programmers. Within hours of Google Maps’ release, programmers had reverse-engineered it, discovering that most of the interactive features relied on simple miniprograms or “scripts” written in JavaScript, one of the Web’s best-known languages for scripting. That meant programmers could write their own JavaScript to manipulate the maps – for example, making their own content, rather than the usual Google search results, appear in the info windows tied to specific map locations.

Almost immediately, programmers started building services atop Google’s map infrastructure. Computer graphics expert Paul Rademacher, for example, launched HousingMaps, a site that pulls real-estate listings off the popular classified-ads site craigslist, uses the addresses of the listed homes and apartments in a given neighborhood to figure out their latitudes and longitudes, and lets users view the properties on a Google map. HousingMaps has no affiliation with craigslist or Google; Rademacher built the hybrid site simply by figuring out how to write coded requests that would grab the appropriate data from the two companies’ public databases. Fortunately, the companies take a mostly benign view of such mash-ups.

Google is so eager to let outside programmers experiment with its mapping platform, in fact, that it released an official API on June 30, meaning hackers would no longer have to waste time on reverse-engineering. That’s led to an even bigger wave of Google Maps creations, from the practical to the disturbing. At ahding.com/cheapgas, you can see gasoline prices from Gasbuddy.com plotted on a Google map, directing you to the lowest-priced pumps in your area. FloridaSexualPredators.com, meanwhile, shows place marks for the homes of every sex offender listed in the Florida Department of Law Enforcement’s public database. Clicking on a place mark brings up an info window with the offender’s name, address, and mug shot.

Geography as Context
To use the Google Maps API, developers must agree not to use the service for commercial purposes, and so far, even Google has refrained from placing ads on Google Maps pages.

But for companies exploring the Internet for the next big business opportunity, the geospatial Web is the equivalent of a virgin continent waiting to be planted with billboards. The attraction is especially great for companies in the search business, for one simple reason: interactive maps have the potential to greatly extend the power of contextual advertising – the engine that drives the search industry and accounts for Google’s ever rising revenues.

Every time you do a search at Google or read a message in your Gmail inbox, you’ll see a different set of ads on the right side of the browser window. This selection isn’t random: each ad relates to a keyword appearing somewhere in your search results or in your mail. Users are more likely to notice ads if they relate to products or services they’re already looking for, as is demonstrated by measurements of the all-important “click-through” rate for contextual ads, which is higher on average than that for other types of ads, such as banner ads. Because search companies charge advertisers by the click, they have a huge incentive to figure out which ads will be most relevant to a user at any given moment and to make sure he or she sees just those ads.

There’s one big drawback to using keywords to tailor ads, however: advertisers lose the opportunity to cater to users’ other interests. Say you’re searching for tickets to The Producers on Broadway. You might also be curious about restaurants near Times Square with early seatings. But today’s search technology can’t hazard such guesses efficiently – and it may never do so. After all, what search engine could divine that a visitor to Chicago is interested in both Cezanne and architecture?

Maps provide a way out of this dilemma. We may be able to communicate instantly with friends halfway around the globe, but we’re still fleshly creatures who must fulfill our basic needs locally. If you’re new to a particular area, looking at a map is the most natural way in the world to search out local services. In this case, the “context” for contextual ads is no longer a list of keywords but a location – meaning that the primary measure of an ad’s relevance to the user is simply proximity, with no fancy psychographic algorithms required.

The developers in Yahoo’s local-services division understood this sooner than Microsoft or Google. In March 2004, they introduced SmartView, a set of buttons alongside a traditional Yahoo map that allows users to highlight points of interest, from spas to sports stadiums. (In an example of custom marketing, the maps can also show special icons for the locations of Carl’s Jr. restaurants, Jeep diesel stations, Intel Centrino-certified Wi-Fi hot spots, and other branded services.)

“It’s cool to see a photo of the Sphinx from a thousand feet up, but we’re focused on understanding people’s key tasks and helping with those,” says Yahoo local-services manager Paul Levine. But Yahoo is also asking outside programmers for help thinking up new ways to deploy Yahoo maps; the company released an API for its mapping service on the same day as Google.

Microsoft, which launched MSN Virtual Earth at the end of July, may appear to be a latecomer to advanced Web mapping. Actually, the company has been in the map business since the early 1990s, offering business-oriented products such as its MapPoint Location Server, which helps companies track shipments or mobile workers, and consumer travel-planning software such as Microsoft Streets and Trips. But Virtual Earth is a different animal, exploiting all the power of the Web-services model to act as something like a “geo-organizer” – a way of managing data intended to complement, and perhaps someday supersede, classic organizational tools such as address books.

The service could become a powerful rival to Yahoo Maps, Google Maps, and even Google Earth. Graphically, it offers satellite views similar to those available from Google Maps. But it also offers some unique features, such as a scratch pad where users can paste in notes about the locations they view. One mouse click lets the user e-mail the scratch pad’s contents to friends or publish them to a blog page on MSN Spaces, Microsoft’s new blog-hosting service.

“You’ll be able to take content from Spaces into Virtual Earth and take content from Virtual Earth into Spaces and share it with whomever you want to share it with,” says Mark Law, lead product manager for MSN Virtual Earth. That will be a boon for Web users, who will gain a new channel for communicating and sharing digital content. And it will be a boon for Microsoft, since every Web page viewed in the process of sharing represents new real estate for contextual ads.

Annotating the Planet
As the big three vie for Web users’ loyalty, they’re likely to introduce more ways for people to import their own data and see it displayed on professional-looking maps. Google Earth Plus, an enhanced subscription version of the program, allows users to upload and view data collected by their GPS units, such as “tracklogs,” series of virtual bread crumbs showing where the user has been.

And other companies are getting into the mix. A program for Nextel GPS camera phones, Trimble Adventure Planner, helps users create online travelogues by uploading photographs and pinning them to the appropriate spots on a Web map.

Siemens, meanwhile, is developing software that will let a GPS-enabled mobile device associate notes with specific coordinates; when someone else with a similarly programmed gadget approaches the coordinates, the note appears on his or her screen. A tourist bureau might “label” a particular spot along San Francisco’s Embarcadero as the site of a fatal duel in August 1879. John Udell, a columnist for InfoWorld, has coined a phrase for this phenomenon: “annotating the planet.”

It’s a trend that the main providers of mapping platforms have every incentive to encourage. After all, as the history of the Web itself has shown, interesting content draws more traffic, which drives more click-throughs. “The world is really dense with information,” says Schuyler Erle. “Access to ubiquitous networking and location-finding services means that we can take that information and make it accessible in the places we are actually in, when we need it, and that allows us to make much more intelligent decisions on the spot, at that time.”

Every page on the Web has a location, in the form of a URL. Now every location can have a Web page – indeed, an infinite stack of them. That may sound like a recipe for information overload. But in fact, it means that navigating both the Web and the real geography around us is about to become a much richer experience, rife with occasions for on-the-spot education and commerce. It means that we will be able to browse the Web – and the virtual earth encompassed within it – simply by walking around.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.