The desire to cluster curves came up recently, and I built the attached protocol to demonstrate using the cluster data component for clustering timer series data.

This protocol also makes a feeble attempt at trying to figure out an optimal number of clusters in the data set, but you can ignore the second half of pipeline one if you are not interested in that.

Reauires the Plate Data Analytics (just for the data reader) and the Data Modeling collections, and makes use of the FLIPR 384 well kinetic read data set.

Enjoy,

Leo