BIG, BOLD and beautiful - Venus and Serena Williams were ranked the most popular players early at the Australian Open, according to data collected using technology that helps fans express how they feel about the players.
Player popularity rankings are displayed on IBM's ''social leaderboard'' on the Australian Open website. The leaderboard tracks Twitter conversations related to each player's hashtag (a tag within Twitter that makes it easy to follow a specific topic), and calculates the percentage of positive and negative sentiment tweeted about players during the Open.
Last year, the leaderboard only tracked the overall volume of tweets. This year, players' rankings are also determined using new social-media analytics and natural-language processing software from IBM to gauge fan sentiment.
This is the 20th anniversary of IBM's partnership with Tennis Australia. ''During the past two decades, we have seen the technology evolve from the launch of the tournament's first website in 1996 to innovative [internet] and big data solutions that bring fans closer to the action and provide deep insight for players, coaches and organisers,'' says IBM marketing specialist Elizabeth O'Brien.
For example, IBM's Social Sentiment Index, new to the Australian Open website this year, tracks the most popular players on a variety of social media. It aggregates and gauges public opinion across hundreds of thousands of posts.
Last week Venus outshone her sister at No.1 on the index while Serena was No.7. Australia's Sam Stosur dropped to ninth place after she lost her match against China's Zheng Jie. This week Serena leapt to the top spot and Venus fell to No. 4 after her third-round loss to Maria Sharapova.
Analysing social sentiment is challenging, O'Brien says. ''It's difficult to work out what's a positive or negative sentiment,'' she says. ''You have to understand the language and slang.'' How does the software distinguish between sarcasm, sincerity and slang - for example, if a player is referred to as ''sick'' or ''bad''? The trick is to look at words surrounding the slang, such as ''awesome'' or ''terrible'', which will indicate whether a sentiment is positive or negative, an IBM technical expert says.
The aim of all this technology is to deepen fans' engagement and enjoyment of the event, says Tennis Australia chief information officer Samir Mahir.
Also, Tennis Australia uses the information to improve its website and the event.
''We are seeing a huge increase in the volume of fan conversation via social media,'' Mahir says.
''Social-media insight has an increasingly important role in how Tennis Australia and other organisations make decisions and engage consumers.''
Sentiment analysis rates the popularity of players and indicates emerging public opinions and trends. From this, Tennis Australia can look at the topics that get fans talking online. The results will help it plan next year's tournament.
IBM has applied similar techniques to gauge sentiment about the Oscars, the US Super Bowl and Major League Baseball, to identify social-media trends and better understand public opinions.
Tracking social media also helps IBM allocate computer resources at the Australian Open. For the two weeks of the tournament, increased website traffic is expected. But by tracking usage of social media, organisations such as Tennis Australia can try to anticipate spikes in activity. At this year's Open, IBM is using predictive software to determine when the official website needs more resources to avoid grinding to a halt in peak periods.
This is the first tennis grand slam event to deploy the predictive provisioning technology, which predicts demand and automatically assigns, ahead of time, the computing power needed to support the website, based on information such as historical data, match schedules, player popularity, and trends gauged from social-media activity.
''With analytics, we can help predict when demand is expected to spike, such as when a crowd favourite goes on court to compete,''
Mahir says. Extra computing resources can come online within three minutes. ''This means fans can expect real-time access to more dynamic content, even during the busiest periods of the competition,'' he says.
1993 Speed serve system begins.
1994 Public touch-screen information kiosks.
1996 Australian Open website launch.
1999 A tennis grand slam first, desktop computer access to match data.
2007 More efficient management of computer workloads.
2008 Real-time statistics presented visually through Slamtracker.
2010 Website hosted by IBM.
2012 Keys to the Match, what players need to do to win a match, introduced on Slamtracker.
2013 Social Sentiment Index and predictive computer resource provisioning technology.
Cynthia Karena attended the Australian Open as a guest of IBM.