Description of the App
TCS SocialSoccer, is an app that digitally reimagined the world’s biggest sporting spectacle, the 2014 FIFA World Cup. TCS SocialSoccer brought fans closer to the matches by using the power of digital forces – social media, big data, mobile and cloud – working in harmony.
The result was a comprehensive social insight into every team and player in the World Cup.
TCS SocialSoccer demonstrated how a combination of digital forces can give powerful and effective analytic solutions, providing meaningful and potentially actionable analytics insights in real time in hands of a user. It uses Perivista, the TCS proprietary tool on big data analytics for sentiment and text analytics. The algorithms are executed in a Hadoop big data cluster on a real time stream of twitter data. Various analytics like sentiment analysis, topic and entity extraction are performed on these tweets and the results are delivered to a mobility web service layer.
Some of the main features of the app are:
• Trending tweets about the favourite team/players
• Twitter trending list of team/players
• Ongoing match analysis with Twitter activity details, tag cloud, sample tweets, live updates of events during the match, e.g. goals and yellow cards
• Locations of where teams/players are popular
TCS SocialSoccer brought fans an ability to relate or understand their counterparts in other parts of the world.
Brazilian fans knew that Neymar was not only the most mentioned player in their country but also in Croatia, their opponents in the tournament opener. However, the reasons for popularity were different; the Brazilians loved him and were supporting him, and the Croatians feared him and were discussing how to stop him.
Fans following the surreal Suarez ‘bite’ controversy on the app would have got a strong indication about his impending move to Barcelona from Liverpool just from the number of mentions he was getting in far-away Catalan capital a week before the news of transfer was officially announced.
Big data analytics
The analytics gave fans an ability to relate or understand their counterparts in other parts of the world. It provided geography-based popularity of players in real time so you could know interesting facts like which player was more popular in a non-participating city like Kuala Lumpur or Mumbai.
Some interesting insights unveiled through the app where:
• Indonesians love to tweet! They were the fourth biggest tweeters during the World Cup, behind the US, UK and Brazil
• Neymar and Thiago’s absence reduced Brazil’s chances considerably in their semi-finals against Germany. Users predicted a German win, even before kick-off, in the app. Negative sentiment peaked for Brazil’s goal keeper Julio Cesar for letting all the goals in. Positive sentiment peaked for German players Neure, Klose and Kroos. Everyone missed Neymar during the match and were tweeting about him
• In the Argentina versus Netherlands semi-final the app revealed that Mascherano is the most popular in the Argentinian cities Buenos Aires and La Plata, even though Messi wears the armband for Argentina. Live match analysis during the penalty shoot-out showed that the Argentinian goal keeper Sergio was the hero of the moment. The app revealed that Netherlands supporters in the twitter world were of the opinion that if goalkeeper Cillessen was replaced with Krul, he would have saved Netherlands in the penalty shootout, in the same way as he did in their match against Costa Rica
As the users started getting more acquainted with the app, they started asking more insightful questions that often had little to do with the game and more to do with analysis. That was where TCS SocialSoccer as an app stood out against hundreds of other World Cup apps that came at the same time.
Data size and impact
TCS SocialSoccer did real time analytics on 96GB of unstructured data (110 million tweets) to provide valuable insights on the world’s biggest sporting event. It scored 22,000+ app downloads in one month. It reached 274k fans on Twitter. Around 9,634 unique users accessed the application in the week 8 July – 15 July 2014. US, UK, Brazil and Indonesia accounted for 54% of the total tweets analysed. Brazil was the most popular team on Twitter. Neymar, David Luiz and Messi tops the popular players chart.
TCS SocialSoccer was at the first position in the new app section and 18th position overall in sports category apps in Google Play. It is rated 4.5 /5 in both Google Play and Apple App Store.
Technology
TCS SocialSoccer was rich with state of the art architectural solution for handling terabytes of unstructured data on a real time basis. The architectural stack was the perfect blend of Kafka, for stream handling and queuing to avoid data loss, Oozie, for workflow scheduling, Hbase, a fault tolerant way of storing huge quantity of unstructured data, and map reduce jobs, to process huge data sets with parallel, distributed algorithms with high performance in a Hadoop ecosystem.
Millions of tweets consumed by Kafka is standardised and structured by the data transformation and cleansing module of the analytics engine. This module extracts the relevant information from the unstructured data, and inserts it into the column oriented database Hbase, that runs on top of HDFS (Hadoop Distributed File System). Hbase supports dynamic creation of columns to which the extracted data can be stored in the required structure. Various analytics like sentiment analysis, topic extraction, NER (Named Entity Recognition) tagging is performed on the relevant tweets and the analytics results are updated on the mobile device on a real time basis. We used Amazon Elastic Load balancer, Clustered Tomcat Servers, lightweight REST Web Service, which was hosted in Asia Pacific region for the application side. The architecture ensured secure transport of analytics data over SSL to the mobility layer.
Considering the number of concurrent users who may use the application platform, the following were implemented at the app server side:
• Caching techniques on the server to improve the server response time and to provide instant response to the end user
• Multiple web servers running behind a load balancer and hosted them on the Amazon Cloud to provide uninterrupted user experience
• Servers were kept in a virtual private cloud to mitigate any risk of denial-of-service attacks
• https was used to ensure secure transactions
• Security reviews and penetration tests were conducted to ensure that application is hack proof
Added to this, caching on the application side was done to provide a seamless end user experience even on jittery network conditions.
The user interface of TCS SocialSoccer incorporates playfulness, upfront presentation of data, eye pleasing subtle background, handholding tips for innovative controls and highlighting of relevant data. This helps to retain the attention of users as well as making them to revisit the app from regularly during the World Cup.
The app is presented in a smooth story flow where users can get a bird’s eye view as well as finer details of the popularity of their favourite teams and players. Another major challenge was the fragmentation of Android devices in terms of screen size and resolution. TCS SocialSoccer displays as intended in MDPI, HDPI, XHDPI, XXHDPI and XXXHDPI devices as well as in iPhone 4 and iPhone 5.
Apple App Store and Google Play links:
• https://play.google.com/store/apps/details?id=com.tcs.socialsoccer&hl=en
• https://itunes.apple.com/us/app/tcs-socialsoccer/id881805106