The Social

A look at the happiness of Facebook's members.

(Credit: Facebook)

Facebook is even more omniscient than you thought: it can now chart the world's collective hopes and dreams and highs and lows--sort of, at least.

The company's data team on Monday launched a trippy new application called the "Gross National Happiness Index." Taking a similar format to its "Lexicon" trend-tracking product, the "GNH" currently displays a graph of data tabulated over the course of the past few years to track the "happiness" of Facebook users based on words picked up in their status messages.

The GNH is currently restricted to United States-based Facebook users--keep in mind, they now represent less than a fourth of the site's 300 million-strong memberships--who have set English as their default language. That will likely change at some undetermined date.

"Earlier this year, data scientists at Facebook started a project to measure the overall mood of people from the United States on Facebook, based on the sentiment expressed in status updates," explained a company blog post by Facebook's Adam Kramer--who is also a Ph.D. student in psychology:

Examples of positive or happy words include "happy," "yay," and "awesome," while negative, or unhappy words, include "sad," "doubt," and "tragic." We also did a brief survey of some Facebook users, which showed that people who use more positive words, relative to the number of negative words, reported higher satisfaction with their lives.

Cupcakes have been known to make people happy. Wonder if Facebook can track that.

(Credit: Caroline McCarthy/CNET)

Holidays like Thanksgiving, Christmas, and Valentine's Day tend to generate spikes in happiness, as do days of historical significance like the election of President Barack Obama. There are notably "sad" days, too, Kramer pointed out, like the double whammy on January 22, 2008, when the Asian stock market took a dive and young actor Heath Ledger was found dead.

In January, Facebook CEO Mark Zuckerberg chatted with blogger Robert Scoble at the World Economic Forum in Davos, Switzerland, and expressed interest in using the staggering amounts of data on the social network to generate a sort of "sentiment engine."

"He said that already, his teams are able to sense when nasty news, like stock prices are headed down, is under way," Scoble wrote at the time. "He also told me that the sentiment engine notices a lot of 'going out' kinds of messages on Friday afternoon and then notices a lot of 'hungover' messages on Saturday morning. He's not sure where that research will lead."

Sound creepy? Facebook doesn't think so. "To protect your privacy, no one at Facebook actually reads the status updates in the process of doing this research," according to Kramer's post. "Instead, our computers do the word counting after all personally identifiable information has been removed."

(Credit: CNET / Josh Lowensohn)

SAN FRANCISCO--The world of Web 2.0 has been criticized for being too much about the nifty ideas and not enough about raking in the dough. So there were likely more than a few sets of ears in the audience on Monday at TechCrunch50 that perked up at the start of the third batch of start-ups presenting: "New Advertising & Monetization Platforms."

The judges included such Silicon Valley marquee names as Google executive Marissa Mayer, industry veteran Marc Andreessen, Sequoia Capital's Roelof Botha, YCombinator founder and investor Paul Graham, and Zappos CEO Tony Hsieh, who sold his company to Amazon this summer.

The first company to present was 5to1, an advertising technology company that tackles the seemingly unsolvable problem of filling up remnant advertising inventory that can't be filled up by premium or direct sales--and which often ends up getting filled by ads that are cheap and irrelevant. 5to1's model lets site owners and publishers fill up their ad inventory as though it's a music playlist.

"What we're talking about here is total control by the publisher," founder and CEO James Heckman said. "No ad is going to show up that you don't like." (He described typical remnant ads as "the dancing fat bellies and the punch-the-monkey ads.")

But some judges were lukewarm on 5to1.

"I think it's a really slick interface but I would just be worried," Tony Hsieh said. "It just seems like a lot of work to have to go through and decide which ads (to run)...my question is how does it scale as a publisher grows."

The next start-up was another advertising platform, DataXu. The focus of DataXu's product is a data dashboard where publishers can buy ads through ad exchanges like Google's and Yahoo's with a highly refined algorithm that promises to show the right ads to the right people at the right time--for example, that news- and sports-related ads get more reception in the morning--and then tracks the success of an ad campaign with all sorts of analytics.

President and CEO Mike Baker called DataXu's offering "rocket science," adding that the underlying technology was actually used by NASA for a Mars mission plan. "What we're doing is actually using machine-learning techniques to take vast amounts of data with a small positive-action subset, which is very consistent with the Internet advertising problem: there are very few clicks and even fewer actions," Baker said, while declining to provide any real trade secrets. "We're applying on top of that the concept of control systems."

SeatGeek co-founder Jack Groetzinger explains how his service can save people money on tickets.

(Credit: CNET / Josh Lowensohn)

Up next was something much more consumer-focused, and that left the audience pretty impressed: SeatGeek, which forecasts concert and sports ticket prices, much like airline price applications like Microsoft's Bing Travel do. Co-founders Jack Groetzinger and Russ D'Souza explained that sometimes ticket prices can drop unexpectedly at the last minute--and sometimes they don't.

The secondary ticket market is around $15 billion, Groetzinger said.

SeatGeek pulls in ticket prices from secondary sellers such as StubHub or Craigslist and then forecasts where they might go based on an algorithm. "We have a system that every day crawls the Internet and pulls in thousands of actual ticket sales," Groetzinger explained. "We're also pulling in other external factors that we know to drive ticket prices." For a baseball game, for example, it can come down to the weather, the starting pitcher, and whether there are popular concerts in town. "Right now we're testing at about 75 to 80 percent accuracy, and that's going up every day as our system learns."

SeatGeek, which says it's already profitable... Read more

The hottest hotspots in New York...for nerds.

(Credit: Sam Lessin)

Just how powerful can the data behind a location-based application be? Extremely.

Earlier this month, the second annual Internet Week New York took place, and Dropio founder and certifiable data nerd Sam Lessin crunched a bunch of numbers based on what his contacts on urban navigation and friend-finding service Foursquare were doing. Lessin was working with a group of fewer than 100 contacts, almost all of whom are involved in the tech and new-media industries (this is the scene that birthed Foursquare and its predecessor Dodgeball, after all), and yet it's a fascinating peek at just how much this kind of data can reveal. He's posted it on his personal file "drop" on Dropio.

Lessin trawled through the data to find what time people checked into coffee shops in the morning (and whether they were doing this earlier or later on a given day), how much people "lost steam" over the course of a party- and conference-filled week, and how much the most popular gatherings actually matched up to the Internet Week New York official schedule. As it turns out, the hottest parties were impromptu, unofficial gatherings at the Standard Hotel and, um, Sing Sing Karaoke.

Obviously, this isn't perfect. Foursquare updates are voluntary, which means that data can't say a thing about what people are doing when they aren't telling the app about it. The presence of an app like Foursquare, too, can also skew social activity: word about the massive impromptu party at the Standard Hotel bar, for example, spread when the Foursquare check-ins started snowballing.

But when you have enough people participating--which, as of yet, Foursquare does not--the critical mass starts to correct some of those issues. It's a fascinating sneak peek at what sort of value this data could have down the road.

What we can also look forward to: pretty infographics, Orwellian privacy concerns. Eek.

For the past few years, Facebook has been flirting with the possibility of supporting the OpenID log-in standard, which calls itself "an open, decentralized, free framework for user-centric digital identity" without actually building support for it.

Now, the massive social network--once famous for its ultra-walled-garden approach to data and user experience--announced Monday that it has become an OpenID "relying party," which basically means that it's started, at last, to deploy support for the standard. Facebook joined the OpenID Foundation in February, even though many considered its Facebook Connect log-in standard to be a proprietary competitor.

But, Monday's announcement indicated, Facebook believes the two can work in tandem.

"We've always let our users express their real world connections," a post on the Facebook blog read. "From the beginning, Facebook users could use their college and workplace identities to establish real world networks. Now, they can use open standards to establish their identities on Facebook."

Most notably, you can now register for a Facebook account with your Gmail account, or can link an existing Facebook account with Gmail or other OpenID-participating services if they support automatic log-in.

"We've always believed that making the user experience as secure, lightweight, and intuitive as possible, which 200 million people can comfortably enjoy and understand, is one of our top priorities," the blog post read. That could be a subtle nod to the fact that OpenID, founded in 2005, has historically been a bit difficult for the non-tech-savvy to comprehend.

Facebook's blog post also said that security concerns have been an issue. In working with the OpenID community, "we shared our experience developing Facebook Connect, where we eventually came up with a design that ensures that users would know that they were providing their login credentials to Facebook, and not some unscrupulous site."

The plus side? Facebook's tests have indicated that if new users can register with an existing Web service account, like Gmail, that they are more likely to stick around.

Considering all the horror stories we hear about photos hosted on Facebook and people, you know, losing their jobs over them, maybe this isn't such a bad thing: The social network acknowledged in a blog post on Sunday evening that 10 percent to 15 percent of the billions of photos it hosts were affected by a storage problem, replaced by a question mark.

But they aren't permanently gone, the post by engineer Evan Priestley insisted. "We've already repaired about one-third of affected photos and expect to complete repairs on another third tonight," he explained. "We still have all your photos because we store them in a way that maintains multiple copies of the data in case of hardware failures like this."

The company still isn't quite sure how the outage happened.

"During an otherwise routine software upgrade on Friday night, we ran into some problems with our photo storage and a few of the hard drives where we store photos apparently failed all at once," Priestley wrote. "We're trying to fully understand what happened, since simultaneous hardware failures like this are rare."

Facebook is no stranger to uptime issues, with minor but noticeable outages hitting the social network as recently as two months ago. This one, however, is different in that it specifically affected the photos hosted on the site, leading some members to grow concerned about mass deletions.

If your photos disappeared over the weekend, they are probably back already. But just to be safe--you really should keep a backup on your hard drive. Really.

Facebook has joined the board of the OpenID Foundation and will host an OpenID Design Summit later this month, according to a post on the social network's developer blog.

This is a bit of a surprise because Facebook has developed its own universal log-in standard, Facebook Connect, which theoretically competes with the nonprofit OpenID standard. It should be noted that Facebook has not yet announced any official plans to make the two compatible, and that just joining the board and hosting an event might not quell the criticism from open-source advocates who say Facebook is still too proprietary in its nature.

Engineer Luke Shepard will be Facebook's representative on the OpenID Foundation board, a corresponding post on the OpenID blog explained, adding that Shepard has been "a huge internal advocate for OpenID" at Facebook. The board also consists of members from Google, IBM, Microsoft, PayPal, VeriSign, and Yahoo as well as seven elected "community" members. Many of the corporate board members joined about a year ago; OpenID creator Brad Fitzpatrick is now employed by Google and has helped to build its OpenSocial developer platform standard.

"Given the popularity and positive user experience of Facebook Connect, we look forward to Facebook working within the community to improve OpenID's usability and reach," the post by David Recordon and Chris Messina read.

Facebook's blog post, written by engineering VP Mike Schroepfer, expressed similar goals. "It is our hope that we can take the success of Facebook Connect and work together with the community to build easy-to-use, safe, open and secure distributed identity frameworks for use across the Web," Schroepfer wrote.

Facebook made a significant portion of its developer platform code open-source last summer.

It's sort of cute, really: blogger Robert Scoble went on a nice snowy stroll with Facebook founder Mark Zuckerberg while the two were in Davos, Switzerland, for the World Economic Forum. Of course, he wrote about it.

Most of what Scoble wrote about his conversation with the young CEO is either information that was out there already or tidbits like the fact that Zuckerberg was teaming up with former British Prime Minister Tony Blair to work the coat check at the World Economic Forum's annual Women's Dinner (aww!), but there was one fairly interesting part: apparently, Facebook is doing some extensive research into tracking user sentiment, and it has a lot of data on hand but isn't yet sure how it will be used.

"Facebook is, he told me, studying 'sentiment' behavior," Scoble wrote. Keep in mind that he's not actually quoting Zuckerberg, so this may be a bit general. "He said that already, his teams are able to sense when nasty news, like stock prices are headed down, is under way. He also told me that the sentiment engine notices a lot of 'going out' kinds of messages on Friday afternoon and then notices a lot of 'hungover' messages on Saturday morning. He's not sure where that research will lead."

We've had a peek at this already with Facebook Lexicon, the social network's trend-tracking search feature. It also sounds a lot like what some people are suggesting as a signature use for Twitter and may explain Zuckerberg's apparent onetime interest in acquiring the microblogging company.

More importantly, this is basically confirmation (via Scoble, of course) that Facebook has significant amounts of intricate data on hand that it hasn't released yet. It may sound creepy, and privacy advocates may be wringing their hands already, but for Facebook, this could be a quick answer to the profitability question.

NOTE: As it turns out, Zuckerberg was participating on Friday in Davos forum, along with Microsoft's Craig Mundie, YouTube's Chad Hurley, Adobe Systems' Shantanu Narayen, and others. The live Webcast has come and gone, but it looks like an archived version might show up eventually on this World Economic Forum Webcast page.

Now, this is kind of neat: Facebook Connect, the sprawling social network's universal-login project, has started to come to desktop software. Namely, it's been hooked up to the Apple photo-management software iPhoto, per a post on the company developer blog.

"We are excited that sharing your photos with the people you care about has become even easier with iLife '09, Apple's new suite of applications that includes iPhoto '09," the post by Facebook platform manager Dave Morin said. "Users of iPhoto '09 can easily share and tag photos from iPhoto directly to Facebook. With help from Facebook Connect, photo tags from iPhoto '09 can be added to Facebook and generate Facebook notifications. Additionally, Mac users can update Facebook News Feed and alert friends anytime they update their websites using Apple's iWeb '09 application."

Basically, this means that if you're a Mac user running the latest edition of its iLife package, which started shipping earlier this week, you can hook up your Facebook account for easy uploading right from iPhoto. If you use the iWeb site creation tool, you can set it up to post a message to your Facebook profile (and your friends' news feeds) if you make some kind of edit. That's pretty similar to what a number of Web-based blogging services have already set up using Facebook's platform.

I haven't actually checked it out yet, so I can't provide a thumbs up or down, but the concept itself is pretty cool. Facebook rolled out its Facebook Connect product, which lets third-party sites (and now desktop apps, apparently) use Facebook usernames and passwords for user accounts, over the second half of last year. The reception, so far, has been generally positive.

What'll be really interesting is to see the further implications of Web-based login standards like Facebook Connect as they're synced up to more desktop applications. Not that you'd really want to share all your Microsoft Word edits in your news feed or anything.

These data portability announcements keep rolling on: On Monday, Google announced that its Google Friend Connect product, which plugs social-networking features into participating sites, is now compatible with Twitter.

So what does this mean? Well, if you go to a site that uses Google Friend Connect, you can opt to use your Twitter credentials to log in to it. Then, as the official Google blog explained, you can then find which of your other Twitter friends are using the same site. Also, you can send out a "tweet" announcing that you've joined up.

Twitter was one of the launch partners for the MySpace Data Availability service, now known as MySpaceID. That has yet to launch, but MySpace has used Google Friend Connect to power the standard, so this could be a sign that it's still on the way.

What's not on Twitter yet? Facebook Connect, the rival log-in product developed by the social network, which rolled out to a full launch on the same day as Google Friend Connect. Rumor has it that Facebook tried to buy Twitter in a failed $500 million deal. There's still no reason to assume Twitter won't integrate Facebook Connect, but for now, it's just Google's alternative.