From: Wade T.Smith (email@example.com)
Date: Thu 28 Nov 2002 - 17:29:33 GMT
Postcards From Planet Google
By JENNIFER 8. LEE
MOUNTAIN VIEW, Calif.
AT Google's squat headquarters off Route 101, visitors sit in the lobby,
transfixed by the words scrolling by on the wall behind the
receptionist's desk: - animación japonese Harry Potter pensées et poèmes
associação brasileira de normas técnicas -
The projected display, called Live Query, shows updated samples of what
people around the world are typing into Google's search engine. The
terms scroll by in English, Chinese, Spanish, Swedish, Japanese, Korean,
French, Dutch, Italian - any of the 86 languages that Google tracks.
- people who shouldn't marry /"she smoked a cigar"/ mr. potatoheads in
long island /pickup lines to get women /auto theft fraud how to -
Stare at Live Query long enough, and you feel that you are watching the
collective consciousness of the world stream by.
Each line represents a thought from someone, somewhere with an Internet
connection. Google collects these queries - 150 million a day from more
than 100 countries - in its databases, updating and storing the computer
logs millisecond by millisecond.
Google is taking snapshots of its users' minds and aggregating them.
Like a flipbook that emerges when successive images are strung together,
the logged data tell a story.
So what is the world thinking about?
Sex, for one thing.
"You can learn to say 'sex' in a lot of different languages by looking
at the logs," said Craig Silverstein, director of technology at Google.
(To keep Live Query G-rated, Google filters out sex-related searches, though less successfully with foreign languages.)
Despite its geographic and ethnic diversity, the world is spending much
of its time thinking about the same things. Country to country, region
to region, day to day and even minute to minute, the same topic areas
bubble to the top: celebrities, current events, products and computer
"It's amazing how similar people are all over the world based on what
they are searching for," said Greg Rae, one of three members of Google's
logs team, which is responsible for building, storing and protecting the
Google's following - it is the most widely used search engine -- has
given Mr. Rae a worldview from his cubicle. Since October 2001, he has
been able to reel off "anthrax" in several languages: milzbrand
(German), carbonchio (Italian), miltvuur (Dutch), antrax (Spanish). He says he can also tell which countries took their recent elections seriously (Brazil and Germany), because of the frenzy of searches. He notes that the globalization of consumer culture means that the most popular brands are far-flung in origin: Nokia, Sony, BMW, Ferrari, Ikea and Microsoft.
Judging from Google's data, some sports events stir interest almost
everywhere: the Tour de France, Wimbledon, the Melbourne Cup horse race
and the World Series were all among the top 10 sports-related searches
last year. It also becomes obvious just how familiar American movies,
music and celebrities are to searchers across the globe. Two years ago,
a Google engineer named Lucas Pereira noticed that searches for Britney
Spears had declined, indicating what he thought must be a decline in her
popularity. From that observation grew Google Zeitgeist, a listing of
the top gaining and declining queries of each week and month.
Glancing over Google Zeitgeist is like taking a trivia test in cultural
literacy: Ulrika Jonsson (a Swedish-born British television host), made
the list recently, as did Irish Travelers (a nomadic ethnic group, one
of whose members was videotaped beating her young daughter in Indiana)
and fentanyl (the narcotic gas used in the Moscow raid to rescue
hostages taken by Chechen rebels in late October).
The long-lasting volume of searches involving her name has made Ms.
Spears something of a benchmark for the logs team. It has helped them
understand how news can cause spikes in searches, as it did when she
broke up with Justin Timberlake.
Google can feel the reverberations of such events, and others of a more
serious nature, immediately.
On Feb. 28, 2001, for example, an earthquake began near Seattle at 10:54
a.m. local time. Within two minutes, earthquake-related searches jumped
to 250 a minute from almost none, with a concentration in the Pacific
Northwest. On Sept. 11, searches for the World Trade Center, Pentagon
and CNN shot up immediately after the attacks. Over the next few days,
Nostradamus became the top search query, fueled by a rumor that
Nostradamus had predicted the trade center's destruction.
But the most trivial events may also register on Google's sensitive
cultural seismic meter.
The logs team came to work one morning to find that "carol brady maiden
name" had surged to the top of the charts.
Curious, they mapped the searches by time of day and found that they
were neatly grouped in five spikes: biggest, small, small, big and
finally, after a long wait, another small blip. Each spike started at 48
minutes after the hour.
As the logs were passed through the office, employees were perplexed.
Why would there be a surge in interest in a character from the 1970's
sitcom "The Brady Bunch"? But the data could only reflect patterns, not
That is a paradox of a Google log: it does not capture social phenomena
per se, but merely the shadows they cast across the Internet.
"The most interesting part is why," said Amit Patel, who has been a
member of the logs team. "You can't interpret it unless you know what
else is going on in the world."
So what had gone on on April 22, 2001?
That night the million-dollar question on the game show "Who Wants to Be
a Millionaire" had been, "What was Carol Brady's maiden name?" Seconds
after the show's host, Regis Philbin, posed the question, thousands
flocked to Google to search for the answer (Tyler), producing four
spikes as the show was broadcast successively in each time zone.
And that last little blip?
"Hawaii," Mr. Patel said.
The precision of the Carol Brady data was eye-opening for some.
"It was like trying an electron microscope for the first time," said
Sergey Brin, who as a graduate student in computer science at Stanford
helped found Google in 1998 and is now its president for technology. "It
was like a moment-by-moment barometer."
Predictably, Google's query data respond to television, movies and
radio. But the mass media also feed off the demands of their audiences.
One of Google's strengths is its predictive power, flagging trends
before they hit the radar of other media.
As such it could be of tremendous value to entertainment companies or
retailers. Google is quiet about what if any plans it has for
commercializing its vast store of query information. "There is
tremendous opportunity with this data," Mr. Silverstein said. "The
challenge is defining what we want to do."
The search engine Lycos, which produces a top 50 list of its most
popular searches, is already exploring potential commercial
opportunities. "There is a lot of interest from marketing people," said
Aaron Schatz, who writes a daily column on trends for Lycos. "They want
to see if their product is appearing. What is the next big thing?"
Google currently does not allow outsiders to gain access to raw data
because of privacy concerns. Searches are logged by time of day,
originating I.P. address (information that can be used to link searches
to a specific computer), and the sites on which the user clicked. People
tell things to search engines that they would never talk about
publicly - Viagra, pregnancy scares, fraud, face lifts. What is
interesting in the aggregate can be seem an invasiion of privacy if
narrowed to an individual.
So, does Google ever get subpoenas for its information?
"Google does not comment on the details of legal matters involving
Google," Mr. Brin responded.
In aggregate form, Google's data can make a stunning presentation. Next
to Mr. Rae's cubicle is the GeoDisplay, a 40-inch screen that gives a
three-dimensional geographical representation of where Google is being
used around the globe. The searches are represented by colored dots
shooting into the atmosphere. The colors - red, yellow, orange - convey
the impression of a globe whose major cities are on fire. The tallest
flames are in New York, Tokyo and the San Francisco Bay Area.
Pinned up next to the GeoDisplay are two charts depicting Google usage
in the United States throughout the day. For searches as a whole, there
is a single peak at 5 p.m. For sex-related searches, there is a second
peak at 11 p.m.
Each country has a distinctive usage pattern. Spain, France and Italy
have a midday lull in Google searches, presumably reflecting leisurely
lunches and relaxation. In Japan, the peak usage is after midnight - an
indication that phone rates for dial-up modems drop at that time.
Google's worldwide scope means that the company can track ideas and
phenomena as they hop from country to country.
Take Las Ketchup, a trio of singing sisters who became a sensation in
Spain last spring with a gibberish song and accompanying knee-knocking
dance similar to the Macarena.
Like a series of waves, Google searches for Las Ketchup undulated
through Europe over the summer and fall, first peaking in Spain, then
Italy, then Germany and France.
"The Ketchup Song (Hey Hah)" has already topped the charts in 18
countries. A ring tone is available for mobile phones. A parody of the
song that mocks Chancellor Gerhard Schröder for raising taxes has raced
to the top of the charts in Germany.
In late summer, Google's logs show, Las Ketchup searches began a strong
upward climb in the United States, Britain and the Netherlands.
Haven't heard of Las Ketchup?
If you haven't, Google predicts you soon will.
Copyright The New York Times Company
This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)
This archive was generated by hypermail 2.1.5 : Thu 28 Nov 2002 - 17:31:57 GMT