Building a Winning March Madness Bracket Using Smart Insights with SAP BusinessObjects Cloud

Building a Winning March Madness Bracket Using Smart Insights with SAP BusinessObjects Cloud

Update: April 4, 2017 - The Final Buzzer

After the final buzzer and some very close games, the SAP DataGenius team finished with 46 of 63 games correct (73%) including correctly predicting the University of North Carolina as National Champions. Big thanks goes out to the team for the analysis and predictions that went into this year’s March Madness bracket and to SAP BusinessObjects Cloud which was easy to use and took the human bias out of the process. 

Until next year, all the best in your analysis!

Update: April 3, 2017 - Championship Game

SAP DataGenius team correctly predicted 45 of 62 games (72%) including the championship game (Gonzaga vs. UNC) for the NCAA March Madness College Basketball tournament. Go Tar Heels!

Who do you have winning tonight?

Update: March 26, 2017 - Sweet Sixteen & Elite Eight Results

This weekend we continued to see strong results in the south region with UNC advancing to the Final Four; which resulted in 100% correct selections for the entire south portion of the bracket. Overall we had 4 out of 8 of the Elite Eight teams correct and selected Gonzaga to reach the Final Four; which resulted in correct selections of 2 out of 4 Final Four teams with the pick for the champion (UNC) still alive. Hope you are enjoying the games!

Update: March 19, 2017 - Round One & Two Results 

In round one we correctly selected 27 out of 32 picks. The results included a few close games that went the other way; Northwestern over Vanderbilt and Michigan over Oklahoma St. There were also several correct upsets we saw in round one analysis including Rhode Island over Creighton, Middle Tenn. over Minnesota, and Michigan St. over Miami.

Then came the upsets in round two which broke our bracket and I’m sure many others with Louisville, Villanova, and Duke all losing out. The DataGenius team correctly selected 10 of 16 Sweet Sixteen team and 2 out of 4 final four selections remain. A bright spot in our bracket was the south region, which was a perfect 12 for 12.

On to the Sweet Sixteen

The team is planning to combine the recent game today with our data model and provide additional analysis leading into next weekend’s Sweet 16 and Elite Eight games. Watch for more updates later in the week. 

How is your bracket doing? Let us know @SAPAnalytics #vizthemadness

---------------------------------------------------------------------------------------------

It’s that time of year college hoops fans! The SAP Data Genius team is diving into the madness using the latest innovations with cloud analytics product SAP BusinessObjects Cloud to breakdown which teams will advance in this year's college basketball tournament. The team used analytic capabilities including; data preparation, modelling, data exploration, planning and what-if analysis, visual storytelling, and automated smart data discovery to uncover hidden patterns for which teams will perform best in the tournament. Let’s have a look at the steps the team used to create the winning bracket. 

Data acquisition

Every year it's a scramble to cobble together interesting and useful data, not that different from how we deal with data and analysis in the business world. This year it was a snap to pull the latest college stats from the web, pulling data from the college basketball stats website kenpom.com.

Linking data

To enrich our data model the team pulled additional insights from the web and added them to the data model. With a few clicks and smart data suggestions provided by BusinessObjects Cloud, we easily joined data together to enhance our model.

Cleansing the data

Once we've created our model, it's a snap to clean and prepare the data the way we want by changing headers, adjusting columns and rows, and removing any messy data. With a few clicks we now have it structured how we like to begin our analysis.

Exploring the data

With the data modeled and cleansed we can switch into data exploration mode and begin asking questions of the data; applying visual charts to better see and understand the data. For example, which teams have the best defense (stopping opponents from scoring) and which have the best offense (scoring points against the opponent)?


Building a story

Now we can begin to create a story based on our analysis, choosing charts and visualizations to support our insights. We can do this manually or have SAP BusinessObjects Cloud automatically do it for us using the “Build my Story” feature.

Find hidden insights and correlations with smart data discovery

Next we can apply automated machine learning algorithms, without needing to know advanced analytical procedures. With a click we can apply machine learning powered by SAP Clea to the data and uncover correlations and insights we may not have considered. For example, which team stats should we consider as indicators of team success?

Planning for the upsets

It goes without saying there will be upsets throughout the tournament; there always is, that's why it's madness! Knowing this, the team decided to plan for the inevitable and use the products planning capabilities to model what-if scenarios to simulate potential outcomes. Based on our data model we’ve added a luck percentage to calculate which teams may have the best chances of pulling off an upset. This allows us to run simulations and model various outcomes based on our selections.

Share and collaborate on our insights

From here, we can bring in the collective knowledge of others on the team by sharing our data story and using collaboration with our DataGenius team around the globe. The team can access the data model and story to apply their own insights.


The SAP DataGenius Bracket

The team is planning to share results as the rounds of the tournament unfold and provide updated analysis, what-if scenarios, and insights based on team performance.


What insights can you uncover?

Do you have what it takes to beat the SAP DataGenius bracket? It’s as easy as 1, 2, 3…

Step 1:

Grab some data:

Google Doc

College basketball data set in Excel

Step 2: 

Try SAP BusinessObjects Cloud

Step 3:

Let the trash talk begin and share your results using #vizthemadness with @SAPAnalytics

Looking forward to seeing what insights you can come up with.

What data project are you working on? The SAP Data Genius team would like to hear about it. Tap into the community of analytics and data experts and join the SAP Data Genius Challenge in 2017, more details to come.

Enjoy the madness!

Khuram (Keyz) Sheikh

There's 8 billion people in the world. I'm just one of them.

6y

Looks awesome Nic!

Like
Reply
Cedric Roussel

Industrial cybernetician - Working on closing the gap between the promise of Industry 4.0 and manufacturing businesses everyday reality | Philosophy student

6y

This is an interesting and funny way to "play" with data analytics. Even if it uses data for a different purpose, it nicely echos the Oakland Atheltics story and how Billy Beane built the success of a "small" major league baseball team using data science. The use of data science is only limited by our creativity.

Randy Bell

database marketing analyst at sas

7y

UNC of course !!!😂

Adem Baykal

Chief Solution Advisor | Customer Advisory | Business Technology Platform | Middle & Eastern Europe at SAP

7y

Nice posting Nic. Thanks for sharing it!

Aaron Graber

Vice President Marketing & Solutions, SAP Business Technology Platform Ecosystem

7y

Good stuff Nic!

To view or add a comment, sign in

Insights from the community

Explore topics