Project Botticelli

In-depth Video: Clustering Data with SQL Server Data Mining

26 June 2013 · 763 views

1-hour 50-minutes and a 12-part demo

Some of you may know that I have been working on a new, in-depth video about clustering, that workhorse of data mining. Clustering is great, but it takes a little while to get used to, because the results it produces are not immediately clear: you just get a bunch of clusters, usefully named Cluster 1, Cluster 2, and so on—and a ton of parameters that describe them. What does all of that mean to you and your data?

Actually, the nature of clustering is not in finding those clusters, which is just the starting point, but in analysing their meaning. I explain how to do this, step by step, in my newest video, Clustering in Depth. The process is relatively simple: you iterate by giving your clusters better-and-better names, while using four different visualisations to try and comprehend their meaning. Once you see it done it becomes second nature—the 12-part demo, which shows it, is the core of this video.

If advanced analytics is important to you, you should learn about Clustering and Decision Trees. That, and the basic concepts, terminology, processes, etc are all nicely covered in our growing online course on Data Mining which also includes a unique learning data set, Happy Cars. There is over 6 hours and 40 minutes of material in 7 modules, but there are more to come! If you are not a Full Access Member yet, consider subscribing for a year, to keep learning at a nice pace, just as the material comes out, every month.

Enjoy clustering!