Google’s Year in Search Is Back and Better Than Ever
Of all the countless things people typed into the Google search box this year, the terror attacks on Charlie Hebdo in January and the city of Paris in November topped the list. That’s not nearly so surprising as the entirely new way Google determined and presented that information.
The findings are the the result of Google’s Year In Search, which in the world of data journalism is a kind of metadata journalism—news about the news that people cared about. Google’s been doing it for 15 years, and this year, it turned the job over to Google News Lab. It created a remarkably comprehensive package, one that used trillions of queries as its data pool. The results are presented as raw search numbers, downloadable datasets, interactive elements, and embeddable components.
True to News Lab’s mission statement (it launched in July), the Year In Search is heavy on visualization tools to help navigate 2015’s major events. “What we started to do this year is treat it as a complex journalism exercise,” says Simon Rogers, data editor at Google News Lab. Rogers, an award-winning data journalist who worked at The Guardian and Twitter before joining Google in March, says his team focused on making the feature more transparent, interactive, and adaptive to local search trends. “In 13 countries, we’ve done a real in-depth focus,” Rogers says. “For everybody else, it’s presented as the most popular global stories, translated into 43 languages.”
Rogers says the year-in-search data is a mix of queries from Google search, Google News, and YouTube. Although Google typically keeps its numbers close to the vest, Year In Search presents the total number of searches across all of those properties. “It’s not really something Google has released before,” Rogers says.
While Google’s Year In Search website offers several options for parsing the data, the best jumping-off point may be the interactive timeline. Mousing over each large circle on the timeline reveals the topic it represents, how many searches it amassed during the year, and when that topic peaked. The Charlie Hebdo and Paris terror attacks were by far the most frequently searched events over the past 12 months.
Google saw nearly 900 million searches on the two attacks, more than twice the queries made for the second most frequently searched topic, the Oscars. Rogers says search patterns during the two attacks provided valuable insight into how news spreads online, even before major news outlets are on the story. With that in mind, the News Lab team created an interactive visualization that showed when certain areas of the world started searching for it, when search queries began in different parts of the world, and the various questions being asked in the queries.
“The event happened at 18 minutes past 9 pm Paris time, and within a couple of minutes, people started searching,” Rogers says. “In Paris, people started asking questions about a minute later. In Berlin, it was a few minutes after that. London after that. And you can see the different questions people were asking, and how it spread around the world. There wasn’t actually a story in the mainstream media for about an hour. This is in the hour before a story is covered by international news media—people are searching for it to figure out what’s going on.”
Clicking the Year In Review timeline reveals slightly different datasets for each event. For example, the page for the Women’s World Cup shows the most popular players in search queries and the comparative search popularity of men’s and women’s soccer. But each topical page also reveals some consistent information, such as country-by-country interest and the most-popular YouTube video associated with a given topic. You can also download packages of data from each page to use for your own projects, and slice it anyway you like. “People can download that data as a CSV, if you’re a data journalist or someone who’s just interested in the data,” Rogers says. “You can do something else with it.”
Of course, as a news hub and aggregator, Google has some serious competition from social networks. Before many of your Twitter and Facebook connections became de-facto editors running their own news bureaus, Google was the only game in town. News organizations could write a search-engine-optimized headline and call it a day.
But even as more immediate (and passive) ways of receiving breaking news proliferate, Google search queries show just how vital googling something remains. People may report and share events via social media platforms like Twitter and Facebook, but they’ll turn to Google when they want to really dig into a story.
Beyond the Year In Search project, Rogers says Google search queries and results can provide unique insights into future events. This year, for example, location-based search queries during general elections in the UK proved to be a better predictor of the outcome than opinion polls.
“In an election year, we look at Google searches for party leaders by the electoral boundaries,” Rogers says. “When I looked at the results, I thought This probably doesn’t mean anything because what it showed was that David Cameron was going to win quite substantially. But the polls were very close, so there was a disconnect. But actually, it was closer to the final result than the polls were.”
Rogers says his team plans to get even more granular in future editions of the Year In Review. He envisions a package where results and visualizations are tailored specifically to where users live, no matter where they live. But first, he wants to see how the world reacts to this year’s package.