Events

The Masters Tournament 2016: Social the Cognitive Way

Share this post:

The Masters Tournament had its origins in 1936 when amateur Bobby Jones found a nursery in East Georgia and exclaimed: “It seemed that this land had been lying here for years just waiting for someone to lay a golf course upon it.” (*)

Since then, the Augusta National Golf Club has transformed into the home of one of the world’s leading sports experiences. This year IBM is celebrating 20 years as the Masters Tournament’s digital partner. Throughout this time, in partnership with The Masters, IBM has introduced many technological innovations, from the tournament’s first website in 1996 to real-time ball tracking in 2015. These transformations and innovations have been implemented not for the sake of innovation, but for the improvement of the experience, always remembering the guiding principles of the club’s founders, Bobby Jones and Clifford Roberts.

This year, the Masters team and IBM examined opportunities to utilize the Watson platform in support of the tournament. An area of particular interest was examining social data to provide insights and analytics for the content that goes into the official social channels and digital platforms. The result is the Cognitive Social Command Center pictured above.

Using Watson to discover social trends in real-time

The many digital properties that touch upon the Masters offer incredible social traffic, trends, and content. Our goal is to leverage that information to support the event; we can also see, in hindsight, what content does well on which platform at what time of day.

While hindsight is incredibly useful, we wanted to help the Masters Team make those changes in flight. The Masters Tournament lasts only four days, so the team’s opportunities to pivot and react are limited. Our team wanted to use Watson to learn how social media is being utilized and, from this knowledge, anticipate trends as they were about to occur and enable the content team to utilize that knowledge.

While retrospectively examining social events is incredibly useful, we wanted to help the Masters Team make those changes in flight.

First, we reviewed platforms on which to develop this system. Watson’s integration and ease of access on Bluemix made the decision simple. This would be one of our team’s first major deployments on Bluemix and we would be using a host of services and buildpacks to support our application.

Detailed below is the overall component diagram for our solution:

CCC Overview

By the way, another member of our team, Aaron Baughman, has written some great stuff on how we design for multiple Bluemix regions to provide enhanced resiliency using the platform. His series begins with Australian Open 2016: Streaming Social Sentiment with Bluemix’s Hybrid Cloud.

Our solution utilizes a number of the Bluemix services; the API and service model was a large part of the decision to use the platform. Often the real complexity of a solution is the integration of components. Bluemix, at least for us, made that integration far less of a concern and more of a decided benefit. A lot of the required integrations were simple “one click” transactions. This ease of integration and the speed at which we could deploy / test / integrate on Bluemix ensured that we could develop our applications in an agile, iterative manner. Additionally, we used the integration with DevOps Services for Track and Plan and source control to coordinate development across the geographies working on the solution.

Solution overview

The solution consists of three major components: social data ingest, cognitive analytical insights, and user experience.

Social Data Ingest

In October 2014, IBM and Twitter announced a partnership for data analytics and insights. Utilizing this dataset, our solution ingests real-time data from Twitter based on PowerTrack rules. This data is received by the Java buildpack and queued for processing in a Bluemix Message Hub queue. The same approach is used for the Facebook integration, the only difference being that the data is polled rather than pushed.

Cognitive Analytical Insights

Once received into the hub, the data is run through Natural Language Classifier, which is being trained to identify data relevant to sport as whole or The Masters Tournament specifically. We then pass those social fragments into Alchemy to identify topics, entities and sentiment. This raw data is then sent to Spark hosted in SoftLayer for aggregation and analysis.

Spark

Finally, once the data has been reviewed it’s sent to a message hub to be pushed into the Cloudant data store.

User Experience

The Cognitive Social Command Center, depicted in the introduction to this post, sits within the Cloudant and Node.js services in Bluemix. We have created a number of views that help further condense and analyze the data, which is then available to API calls sent from a frontend.

We have been really pleased with the Cloudant UI’s ability to help us quickly build and debug views. Our Node.js application uses the Nano and Cloudant npm modules to connect to and return the data. We also implemented a Swagger RESTful service to speed up service definition and help validate requests into and out of the system.

Conclusion

What has impressed us most is the speed at which we can develop and deploy. Even as the application went online this week, we have continued using Cloud Foundry tools to deploy iterative updates, sometimes multiple times a day, without the end users even realizing it. To put it simply, using Bluemix is both simple and expedient for highly iterative prototyping and development.

(*) Bobby Jones from “Golf Is My Game”
(**) Our thanks to Darren Shaw (@shawdm) and Mark Crawley (@crawlem) who were integral to the build.

More Events stories
April 29, 2019

Transforming Customer Experiences with AI Services (Part 1)

This is an experience from a recent customer engagement on transcribing customer conversations using IBM Watson AI services.

Continue reading

April 26, 2019

Updated Tutorial: Database-Driven Chatbot

The tutorial on how to build a database-driven chatbot has been updated. It's now simpler to deploy and and offers more options—Slack, Facebook Messenger, Wordpress, and more.

Continue reading

April 23, 2019

Announcing the Deprecation of the Watson Machine Learning JSON Token Authentication Service

We’d like to inform you about the deprecation of the Watson Machine Learning JSON Token Authentication service. This method of authentication will be retired on May 30, 2019.

Continue reading