Sports Performance Platform: Under the Hood with the Seattle Seahawks

Predictive Analytics for Professional Sports Teams

Earlier this year, we announced the availability of the Microsoft Sports Performance Platform – a predictive analytics platform for athletes and teams. The level of interest in the platform has been amazing to see and has sparked many ongoing and fruitful conversations with customers worldwide about what’s possible.

As a part of the ongoing engagements we’ve been doing with a number of sports teams, this week we have the great opportunity to share our work with the Seattle Seahawks through local broadcast coverage, an awesome transform blog post with video, an example Power BI dashboard on the Power BI Solution Gallery, an inspirational video about the rehabilitation of DeShawn Shead, and a blog post from Microsoft executive Chris Weber. The work with the Seahawks has been really exciting for us to engage in (not to mention we’re helping our hometown team), and similar to our engagements with professional sports organizations like Benfica, Real Madrid, Cricket Australia, and Seattle Reign, our engineering team has been working side by side with the Seahawks Sports Science staff to break new ground.

The result of this work includes a customizable platform that helps the Sports Science staff support athlete performance and reduce players time off the field resulting from injuries, among other supported scenarios. Below will give you a glimpse in just some of the ways the Seahawks are capitalizing on the Sports Performance Platform.

The below screenshot provides a view of a standard players Wellness questionnaire. The dashboard collects wellness data from a player through a mounted kiosk (Surface app) outside of the Seahawks locker room. The data that results from this questionnaire helps provide insights into different aspects of individual athlete and team wellness, which the Seahawks staff use to help prepare for game day.

While the Surface app provides a way to enter (and visualize) data, Power BI provides a rich platform for creating custom insights and analyses. For example, the below report (a facsimile of what the Seahawks use), a Readiness report, uses data such as the Wellness survey data and begins to analyze it in different ways. The Readiness report provides a comparative view (time-series based) by the ID of a player that shows where there may be important deltas to investigate—e.g. deltas where there is low sleep quality and higher levels of soreness (so the need for improved recovery) or deltas between mood and stress (so extraneous factors to the game are explored and put to rest before the team gets on the field). 

To explore the interactive report, visit the Power BI dashboard on the Power BI Solution Gallery.

These are just a couple of examples of how the Seahawks are using the Sports Performance Platform and Microsoft technology more broadly to support player performance, prevent injuries and keep players healthy for game day. 

Predictive Analytics and Sports Performance Platform

The goal of the Sports Performance Platform is to continuously learn and adapt using data and more broadly evolve the Sports industry through a more predictive analytical lens. The general surface area of sports analytics is dynamic and broad—there is a range of problem spaces that can be tackled with small to large software solutions as answers to those problems (which cut across many different areas of sports—e.g. business, fan engagement, performance, etc.). For example, if you’re interested in using sports data feeds to run some predictive analyses for athlete and/or team performance, this can be done, for example, in short order with a data pipeline, a lean set of Cloud-based services and database, a responsive user experience, and a small set of predictive models that can show where trends are heading.

That said, for customers using the Sports Performance Platform, things begin to get interesting when you broaden the type, diversity and structure of data within the data pipeline, run from proprietary-based data to include broader objective statistical data and business data; integrate orchestration and machine-learning engines and workflow; attach a deep reporting and analysis platform; integrate the predictive analytics performance scenario programmatically into other business and systems within the team/organization; and integrate a surround of business and productivity software and devices around the fundamental performance analytics scenario. Wrap this with a full surround of subscription Cloud services and devices, and you’ve got a top-to-bottom solution. Further, you’ve got a jump-off point to integrate other areas of the Microsoft stack, such as Cognitive Services, Azure Data Factory, and so on. 

An Architectural Overview of V1

While we see and are planning for future versions that do integrate more sophisticated technology and approaches, the Sports Performance Platform V1 architectural approach was lean and meant to incubate and simply prove an answer to a technical challenge and take full advantage of existing Cloud services and Microsoft products—e.g., Power BI and Azure Machine Learning (ML). Focusing on the API layer down, we created a rich set of REST APIs that ingest and move data into a Cloud-based data lake. With an authorization wrapper (supporting facial recognition), we plugged an extensible Surface app to the APIs and also integrated a Data Loader tool (that loads and appends disparate data into the Cloud-based data lake). With APIs and a Data Loader tool, customers can ingest multiple points of data into their data lake. We also shipped a Power BI template along with some sample data, which provides a starting point for customers to explore, analyze and report out on data—Power BI is a very powerful analytical platform unto itself. With all this in hand, our V1 provides customers with a good entry point to ingest, analyze, build predictive models, and report on data.

V1 is a beginning point for our customers. For example, using Azure ML capabilities built into Microsoft Azure (or even deploy models that you’ve built using open-source technologies), customers can then build their own machine-learning models and then deploy them as web services. This puts the power of the Sports Performance Platform in our customers hands, as they can integrate these machine-learning services into other parts of the organization.

For technology partners like the Seahawks, they have a great internal staff that are conversant with some of the more modern predictive analytical techniques, which makes deploying and leveraging the Sports Performance Platform a powerful proposition. They’ve been able to take advantage of what we’ve built to purpose-build their own proprietary ‘secret sauce.’ We see many more teams and organizations operating like this, leading us to observe high value IP deriving from the predictive models that are specific to an athlete, team or scenario. There of course is a value-chain of technology, experience and skill that gets them to their endpoint—and the strategic conflation of these three areas is of paramount importance

Where to From Here?

The engineering team has already begun working on the Sports Performance Platform V2. There are many requirements that we have on our backlog from V1, and we are exploring a broader set of scenarios and view of the future for the Sports Performance Platform—and we are already excited by the work that is underway. On our near-term backlog are items like code consolidation and automated deployment for V2—to ensure we update our open-source repository and current partners with an easier-to-deploy-and-use code-base. An example of what this looks like is shown in the following screenshot, where our new deployment scripts deploy the key Cloud assets associated with the Sports Performance Platform and deploy them directly into your Azure subscription.     

Beyond these smaller backlog items, we have a broader set of cross-cutting scenarios beyond performance we're working on.

We anticipate the Sports Performance Platform V2 approach—across architecture, customer integration and partner—will be a big step forward for us and we anticipate it will continue to revolutionize the Sports industry.