IBM Cognos Analytics

How visualizing data in Cognos Analytics helped fill out my bracket this year

Blog Home > How visualizing data in Cognos Analytics helped fill out my bracket this year

How visualizing data in Cognos Analytics helped fill out my bracket this year


Reading Time: 4 minutes

Editor’s note: Last week, I shared my bracket picks after using Watson Analytics and, gosh darn it, Villanova went and ruined everything. It was somewhat satisfying to know that I had picked Duke’s early exit; unfortunately, they lost to South Caroline and not Marquette. This week, our guest poster, Eliot Botwick, shares what I consider another smart bracket move: using Cognos Analytics for help visualizing his picks.

Happy March everyone! One of the most beloved and despised sporting events of the year is upon us, and it’s been as exciting as anyone could have hoped for. I’ve always loved the tradition of filling out brackets, choosing one team to win all the way from the round of 64. However, I haven’t had much success in the past by picking upsets on whims and going off my gut feeling.

So, this year I took a new approach.

Visualizing data to make picks

Sitting at the 85th-percentile of all brackets in the country going into the second weekend, this new approach is working well for me so far. To begin, I obtained a set of predictive data for this year’s tournament from a statistics site, fivethirtyeight.com, that assigns each team odds of winning for each round. In raw form, the data was extremely difficult to digest so I uploaded to Cognos Analytics – IBM’s business intelligence platform. In a few clicks I was able to filter out columns that I did not consider relevant and prepare the data for analysis quickly and easily.

In Cognos, I made a dashboard with various visualizations describing the likelihood of a team advancing in any round. Adding breakdowns by region, seed, and other factors made the data easy to understand.

bracket odds by rounds

Here, a treemap shows the chances teams in the stacked south region have of advancing in the middle rounds of the tournament.

Picking upsets

Every year minds are blown and brackets are busted as favorites are knocked off by unlikely opponents. Already this year, overall favorite Villanova lost to 8-seed Wisconsin and no. 7 South Carolina beat 2nd-seeded Duke. Only one out of four 6 vs. 11 matchups was taken by the 6-seed. We know upsets are coming, but predicting which to choose feels arbitrary without data backing one’s decision.

bracket odds 64

Again looking at the South region, Wichita State and Middle Tennessee (seeds 10 and 12 respectively) both had notably high odds of upsetting their opponents in the round of 64 – and both did.

My bracket

I filled out two this year: one bracket the old fashioned way and one with the help of Cognos Analytics. Going into the weekend of 16 teams, the Cognos-inspired bracket is doing considerably better (85th-percentile versus 65th percentile of all brackets). You can check out my picks below –

bracket cognosI couldn’t help but pick my alma mater University of Virginia making a deep run (go hoos). I took the Kansas Jayhawks to cut the nets this year, mostly because their path to the final four looked easier than most.

Above, Kansas and Purdue’s odds in each round are highlighted – note the large dip in Purdue’s odds (lime green) going into the Sweet Sixteen matchup with Kansas while the Jayhawks’ odds (purple) only diminish gradually through the final four.

Takeaways

The men’s college basketball tournament poses a fun challenge every year for millions of basketball fans, statisticians, and of course the players and coaches. Filling out a bracket may be just for fun but we talk about living in an age of information, and harnessing that information to make smarter decisions. Businesses and industries are changing and maybe the way you fill your bracket out should too.

I think a lot of users would be surprised at how easy Cognos Analytics is to use. The program is certainly not limited to the IT user – it has many self-service features like the ability to type in “round 2 win percentage by region” and instantly generate a great looking plot of each teams chances of winning in the second round broken down by region.

Easily building compelling visualizations helped me make my data-driven decision, it could help you and your business, too.

Have more questions about Cognos Analytics or want to see a demo that walks through the dashboard? Message me at Elliott.Botwick@ibm.com!

Thanks for reading and good luck in your bracket pools!

Take your analytics pick

Watson Analytics and Cognos Analytics are both winners – on the court or in your business. Try either or both today. Here’s where you can go to sign up for the free trials:

(Originally posted on LinkedIn.)

About the blogger

Elliott Botwick is an IBM Predictive Analytics Architect based in the greater Atlanta area serving the financial services industry on the East coast. He attended the University of Virginia, which has a storied basketball tradition, where he earned a Bachelor of Applied Science in Systems Engineering.

Leave a Comment
0 Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More IBM Cognos Analytics Stories

IBM Cognos Analytics

Mohammed Omar Khan

IBM and PMsquare partner to help you see the bigger picture with Thrive!

From data prep and discovery to data visualization and collaboration, AI drives the all new IBM Cognos Analytics! Our goal is to help you hit the ...

IBM Cognos Analytics

Letha Wicker

3 Cognos Analytics Success Scenarios

It’s all very well to list speeds and feeds, pretty screenshots and teaser videos. However, the true power of the all-new IBM Cognos Analytics only ...

IBM Cognos Analytics

Timothy Walker

Announcing IBM Cognos Analytics for IBM Cloud Private for Data

IBM Cloud Private for Data is a platform that continually improves itself; with the addition of IBM Cognos Analytics, that drive only accelerates.