So last week I used SAP Predictive Analytics to cluster the teams – see NCAA Teams Automatic Clustering with SAP Predictive Analytics 2.0
How did it match up with the teams selected? I used the output of Predictive Analytics to analyze the results:
The above shows that cluster 8 has the most teams selected, with cluster 10 a close second.
Cluster 8 has the heavyweight teams like Duke, Kentucky, Virginia. I can see that when I drill down. All the 1 seeds in the tournament were in Cluster 8.
I was not surprised when my co-worker told me that Georgia State beat Baylor, because they were in Cluster 8. Baylor was in Cluster 10. Was Dayton really a surprise winner over Providence last night? Not to me, as they were in cluster 8.
The surprise was that Iowa State lost in the first round; they were in cluster 8.
Cluster 8 had the most wins in the first round – 14 out of 19 teams won their first round games as shown above. Cluster 10 teams won 10 first round games out of 18 teams.
Cluster 5 teams experiences the most losses in the first round.
Cluster 4 only has one win. Guess who?
Feel free to review my cluster results here on Lumira Cloud.