Welcome!

@CloudExpo Authors: Liz McMillan, Zakia Bouachraoui, Yeshim Deniz, Pat Romanski, Elizabeth White

Related Topics: @DXWorldExpo, @CloudExpo, @ThingsExpo

@DXWorldExpo: Blog Feed Post

Golden State Warriors Analytics Exercise | @BigDataExpo #BigData #Analytics

Identifying and quantifying variables that might be better predictors of performance

For a recent University of San Francisco MBA class, I wanted to put my students in a challenging situation where they would be forced to make difficult data science trade-offs between gathering data, preparing the data and performing the actual analysis.

The purpose of the exercise was to test their ability to “think like a data scientist” with respect to identifying and quantifying variables that might be better predictors of performance. The exercise would require them to:

  • Set up a basic analytic environment
  • Gather and organize different data sources
  • Explore the data using different visualization techniques
  • Create and test composite metrics by grouping and transforming base metrics
  • Create a score or analytic model that supports their recommendations

I gave them the links to 10 Warrior games (5 regulation wins, 3 overtime losses and 2 regulation losses) as their starting data set.

I then put them in a time boxed situation (spend no more than 5 hours on the exercise) with the following scenario:

You have been hired by the Golden State Warriors coaching staff to review game performance data to identify and quantify metrics that predict a Warriors victory

Here were the key deliverables for the exercise:

  1. I wanted a single, easy-to-understand slide with in-game and/or player recommendations.
  2. I wanted a break out of how they spent their 5 hours across the following categories:
  • Setting up your analytic environment
  • Gathering and organizing the data
  • Visualizing and analyzing the data
  • Creating the analytic models and recommendations
  1. Finally, I wanted back-up information (data, visualizations and analytics) in order to defend their in-game and/or player recommendations.

Exercise Learnings
Here is what we learned from the exercise:

Lesson #1: It’s difficult to not spend too much time gathering and cleansing data. On average, the teams spent 50% to 80% of their time gathering and preparing the data. That only left 10% to 20% of their time for the actual analysis. It’s really hard to know when “good enough” is really “good enough” when it comes to gathering and preparing the data.

Lesson #2: Quick and dirty visualizations are critical in understanding what is happening in the data and establishing hypotheses to be tested. For example, the data visualization in Figure 1 quickly highlighted the importance of offensive rebounds and three-point shooting percentage in the Warriors’ overtime losses.

Figure 1: Use Quick Data Visualizations to Establish Hypotheses to Test

Lesson #3: Different teams came up with different sets of predictive variables. Team #1 came up with Total Rebounds, Three-Point Shooting %, Fast Break Points and Technical Fouls as the best predictors of performance. They tested a hypothesis that the more “aggressive” the Warriors played (as indicated by rebounding, fast break points and technical fouls), the more likely they were to win (see Figure 2).

Figure 2: Testing Potential Predictive Variables

Team #2 came up with the variables of Steals, Field Goal Percentage and Assists as the best predictors of performance (see Figure 3).

Figure 3: ANOVA Table for Team #2

Team #2 then tested their analytic models against two upcoming games: New Orleans and Houston. Their model accurately predicted not only the wins, but the margin of victory fell within their predicted ranges. For example in the game against New Orleans, their model predicted a win by 21 to 30 points, in which the Warriors actually won by 22 (see Figure 4).

Figure 4: Predicting Warriors versus New Orleans Winner

And then in the Houston game, their model predicted a win by 0 to 10 points (where 0 indicated an overtime game), and the Warriors actually won that game by 9 points (see Figure 5).

Figure 5: Predicting Warriors versus Houston Winner

I think I’m taking Team #2 with me next time I go to Vegas!

By the way, in case you want to run the exercise yourself, Appendix A lists the data sources that the teams used for the exercise. But be sure to operate under the same 5-hour constraint!

Summary
A few other learnings came out of the exercise, which I think are incredibly valuable for both new as well as experienced data scientists:

  • Don’t spend too much time trying to set up the perfect analytic environment. Sometimes a simple analytic environment (spreadsheet) can yield consider insights with little effort.
  • Start with small data sets (10 to 20GB). That way you’ll spend more time visualizing and analyzing the data and less time trying to gather and prepare the data. You’ll be able to develop and test hypotheses much more quickly with the smaller data sets running on your laptop, which one can stress test later using the full data set.
  • Make sure that your data science team collaborates closely with business subject matter experts. The teams that struggled in the exercise were the teams that didn’t have anyone who understood the game of basketball (not sure how that’s even possible, but oh well).

One of the many reasons why I love teaching is the ability to work with students who don’t yet know what they can’t accomplish. In their eyes, everything is possible. Their fresh perspectives can yield all sorts of learnings, and not just for them. And yes, you can teach an old dog like me new tricks!

Appendix A:  Exercise Data Sources
Extract “Team Stats” from the Warriors Game Results website: http://www.espn.com/nba/team/schedule/_/name/gs.  Listed below is a cross-section of games from which you may want to use to start your analysis.

Wins

Rockets 1/20/17: http://www.espn.com/nba/matchup?gameId=400900067

Thunder 1/18/17: http://www.espn.com/nba/matchup?gameId=400900055

Cavaliers 1/16/17: http://www.espn.com/nba/matchup?gameId=400900040

Raptors 11/16/16: http://www.espn.com/nba/matchup?gameId=400899615

Trailblazers 1/2/17:  http://www.espn.com/nba/matchup?gameId=400900139

Overtime (Losses)

Houston 12/1/16: http://www.espn.com/nba/matchup?gameId=400899436

Grizzles 1/6/17: http://www.espn.com/nba/matchup?gameId=400899971

Sacramento 2/4/17: http://www.espn.com/nba/matchup?gameId=400900169

Losses

Spurs 10/25/16: http://www.espn.com/nba/boxscore?gameId=400899377

Lakers 11/4/16: http://www.espn.com/nba/matchup?gameId=400899528

Cavaliers 12/25/16: http://www.espn.com/nba/matchup?gameId=400899899

Note: You are welcome to gather team and/or individual stats from any other games or websites that you wish.

The post Golden State Warriors Analytics Exercise appeared first on InFocus Blog | Dell EMC Services.

Read the original blog entry...

More Stories By William Schmarzo

Bill Schmarzo, author of “Big Data: Understanding How Data Powers Big Business” and “Big Data MBA: Driving Business Strategies with Data Science”, is responsible for setting strategy and defining the Big Data service offerings for Hitachi Vantara as CTO, IoT and Analytics.

Previously, as a CTO within Dell EMC’s 2,000+ person consulting organization, he works with organizations to identify where and how to start their big data journeys. He’s written white papers, is an avid blogger and is a frequent speaker on the use of Big Data and data science to power an organization’s key business initiatives. He is a University of San Francisco School of Management (SOM) Executive Fellow where he teaches the “Big Data MBA” course. Bill also just completed a research paper on “Determining The Economic Value of Data”. Onalytica recently ranked Bill as #4 Big Data Influencer worldwide.

Bill has over three decades of experience in data warehousing, BI and analytics. Bill authored the Vision Workshop methodology that links an organization’s strategic business initiatives with their supporting data and analytic requirements. Bill serves on the City of San Jose’s Technology Innovation Board, and on the faculties of The Data Warehouse Institute and Strata.

Previously, Bill was vice president of Analytics at Yahoo where he was responsible for the development of Yahoo’s Advertiser and Website analytics products, including the delivery of “actionable insights” through a holistic user experience. Before that, Bill oversaw the Analytic Applications business unit at Business Objects, including the development, marketing and sales of their industry-defining analytic applications.

Bill holds a Masters Business Administration from University of Iowa and a Bachelor of Science degree in Mathematics, Computer Science and Business Administration from Coe College.

CloudEXPO Stories
With more than 30 Kubernetes solutions in the marketplace, it's tempting to think Kubernetes and the vendor ecosystem has solved the problem of operationalizing containers at scale or of automatically managing the elasticity of the underlying infrastructure that these solutions need to be truly scalable. Far from it. There are at least six major pain points that companies experience when they try to deploy and run Kubernetes in their complex environments. In this presentation, the speaker will detail these pain points and explain how cloud can address them.
The deluge of IoT sensor data collected from connected devices and the powerful AI required to make that data actionable are giving rise to a hybrid ecosystem in which cloud, on-prem and edge processes become interweaved. Attendees will learn how emerging composable infrastructure solutions deliver the adaptive architecture needed to manage this new data reality. Machine learning algorithms can better anticipate data storms and automate resources to support surges, including fully scalable GPU-centric compute for the most data-intensive applications. Hyperconverged systems already in place can be revitalized with vendor-agnostic, PCIe-deployed, disaggregated approach to composable, maximizing the value of previous investments.
When building large, cloud-based applications that operate at a high scale, it's important to maintain a high availability and resilience to failures. In order to do that, you must be tolerant of failures, even in light of failures in other areas of your application. "Fly two mistakes high" is an old adage in the radio control airplane hobby. It means, fly high enough so that if you make a mistake, you can continue flying with room to still make mistakes. In his session at 18th Cloud Expo, Lee Atchison, Principal Cloud Architect and Advocate at New Relic, discussed how this same philosophy can be applied to highly scaled applications, and can dramatically increase your resilience to failure.
Machine learning has taken residence at our cities' cores and now we can finally have "smart cities." Cities are a collection of buildings made to provide the structure and safety necessary for people to function, create and survive. Buildings are a pool of ever-changing performance data from large automated systems such as heating and cooling to the people that live and work within them. Through machine learning, buildings can optimize performance, reduce costs, and improve occupant comfort by sharing information within the building and with outside city infrastructure via real time shared cloud capabilities.
As Cybric's Chief Technology Officer, Mike D. Kail is responsible for the strategic vision and technical direction of the platform. Prior to founding Cybric, Mike was Yahoo's CIO and SVP of Infrastructure, where he led the IT and Data Center functions for the company. He has more than 24 years of IT Operations experience with a focus on highly-scalable architectures.