Table pivot without aggregation

Version 1.1 2021-05-14

This workflow allows users to perform pivoting without data aggregation.

While a pivot table is a useful tool in reorganizing and summarising data in a table, there are instances whereby it would be beneficial to retain the individual replicate raw data. For example, when there is a need to evaluate individual trends in each replicate (e.g. time series, linear regression, concentration-response curve etc.).

This workflow shows the difference between the usual pivot approach (upper branch) and the proposed pivoting without aggregation (lower branch).

Usual pivot approach (upper branch in this workflow):
Column A (with aggregation method, e.g. mean)
Aggregative value 1
Aggregative value 2
Aggregative value 3
Aggregative value 4
Aggregative value 5

Pivoting without aggregation (lower branch of this workflow):
Column A (individual replicate #1), Column B (individual replicate #2), Column C (individual replicate #3),
Individual value 1a, Individual value 1b, Individual value 1c
Individual value 2a, Individual value 2b, Individual value 2c
Individual value 3a, Individual value 3b, Individual value 3c
Individual value 4a, Individual value 4b, Individual value 4c
Individual value 5a, Individual value 5b, Individual value 5c

To perform pivoting without aggregation, the List aggregation method is used to group all the replicates without summarising them into a summary statistics such as mean, median etc. The list is then split into individual columns (which should align to the respective replicates), and the appropriate column header is applied to each column using dynamic flow variables.

As always, it is hoped that this workflow can be useful for those who require a dynamic approach to pivoting without losing its individual replicate raw data. Any questions and/or feedback will be much appreciated!