Advanced Analytics & Wimbledon
During the two weeks of The Championships, Wimbledon engages with tennis fans using both traditional and social media. Their goal: to drive and be part of the conversation that happens around the event. As part of IBM’s relationship with the All England Lawn Tennis & Croquet Club (AELTC), the Emerging Technology team used technology to provide novel information and statistics in a timely and rapid manner–basically before anyone else could.
Working with teams from IBM Watson, IBM InfoSphere Streams, and IBM Research, we delivered three different projects to help Wimbledon push technology to the edge of the baseline: Streams Analytics, Social Media Analytics and Watson. Let’s take a look at each of these projects:
The AELTC at Wimbledon wanted to improve their ability to engage with tennis fans by providing near-realtime alerts for key events in matches. Working with the AELTC content team Emerging Technology identified the events of interest and then developed realtime analytics, using IBM InfoSphere Streams, to identify those key events during the Championships. As a result of these alerts the AELTC were able to publish timely updates to the fans using Twitter and the Wimbledon website.
IBM captured data at courtside — as play occurs in real time — for every point played in every match. Streams analyzed the flow of that data to glean any key events, deviations in play, patterns of play and characteristics of player performance. First superfluous data was filtered out so that only relevant data was processed. Then the valuable data was analyzed against information that was stored in Wimbledon’s database containing match and player statistics that are maintained throughout the tournament.
The analytics pushed insights as soon as it was generated in the context of the matches to those members of the Wimbledon team who needed it. The goal was to notify team members first with the information on key elements of interest. For example, Roger Federer hit his 2,500th winner of his Wimbledon career at this year’s tournament, and Wimbledon officials were able to immediately respond to that milestone for fan engagement using Twitter.
Capturing the action on the tennis court to analyze data flow in real time and derive relevant data in context enabled the Wimbledon team to make on-the-spot decisions about content to publish to their customers, the fans.
The analytics developed by Emerging Technology are currently being deployed for the US Open tennis tournament, extending the value of the work of ET and demonstrating yet another way InfoSphere Streams technology can be applied to deliver insights that allowed Wimbledon to provide a more immersive experience.
Social Media Analytics
Wimbledon’s social media team wanted to respond to the conversation the fans were having on social media and direct their content to match their interests. The Emerging Technology team developed a system to identify topics, players, and events that held the public’s attention.
IBM pulled a live feed from Twitter. At this stage the Tweets were crudely filtered so as to identify messages that were genuinely about Wimbledon, and discarding extraneous noise (which there was a lot of). Key to the success of this effort was our partnership with the Watson Explorer (WE) team, leveraging their technology to filter the Twitter stream with text analytics to identify content that was actually about Wimbledon. With this filtered stream, more in-depth analysis took place, again using WE, this time to identify things like topics, players, celebrities, locations and sentiment.
Using expertise from IBM Research, we were able to implement sophisticated algorithms to discover social influence by understanding social network interactions and relationships to determine key influencers relating to Wimbledon. The analysis was done in realtime and the results were rendered in a Command Centre dashboard available to Wimbledon’s social media team. That gave them a realtime view of what people were talking about in relation to Wimbledon. Using this information the social media team responded and produced new content, further driving engagement with Wimbledon’s social media profiles.
Journalists creating content about Wimbledon often rely on help from AELTC and IBM with research. This could be as simple as identifying a statistic or a trend about a player or perhaps digging up background historical information about the tournament. In the past Wimbledon has relied on data scientists writing queries or librarians to search the document archives.
For 2015, the Emerging Technology team developed a Watson-based solution that could answer both statistical questions such as, “What is Andy Murray’s average serve speed in semi-finals?” and general knowledge queries such as “why do the players wear white?”–both posed in natural language.
Building on the work by the Watson team which created the Watson Engagement Advisor (WEA), we trained WEA to use Wimbledon’s document archives to take care of the general knowledge questions.
Hudson, a component built using recent research advances in Controlled English (and developed by the Emerging Technologies team in the UK in partnership with IBM Research), was used to tackle the statistical questions. It provided a fixed model of the world that data sources were then mapped on to. For instance, the model may have described a serve as something that a tennis player does, which has a speed. It then represented the model and data in structured form of English, meaning that it was possible to map questions posed in everyday normal English on to its structured English model of the world, and thus connected questions to data sources, and ultimately the answer.
The two components were integrated in to a single user interface, giving the media a single point to ask any questions about Wimbledon.
All of these systems leveraged cloud infrastructure capabilities from IBM, by leveraging SoftLayer.
Working closely with the Wimbledon team, as well as IBM teams from Watson, InfoSphere Streams, and IBM Research, Emerging Technologies was able to developed technology which provided capabilities including realtime streaming analytics, social media engagement, and tools for the media using Watson technologies, creating a new level of sophistication for the world’s oldest and most prestigious tennis tournament in the world.