Just a few weeks ago, I only knew the term “data mining“ from my previous research into Bitcoins ( = complicated stuff!!)
Imaging my surprise when I found out that data mining is a lot closer to home:
Techniques used for discovering patterns in large data sets.
Segmentation or clustering is one of several data mining techniques.
If you run a clustering analysis on your customer information, the output will be a segmentation based on your customers natural occurring behaviour. Your criteria are not predefined but a result of the analysis. Interesting approach!
Traditional segmentation was typically based on predefined demographic variables. But the amount of data available was a lot smaller only a decade ago, so the whole segmentation exercise was really different.
A simple example
Tableau has a built-in clustering analysis that I used for a recent Makeover Monday assignment (dataset from European Commission).
The clustering analysis does not simply split the dataset in three equal parts, but finds common behaviour a basis for the grouping.