This package is inspired by the Managing many models-talk by Hadley Wickham. The intuition is to fit arbitrary models to a signal. Those models are used to generate predictions. An (error)-metric gets computed, based on the predicted and the original value of the signal. Using dplyr, relevant measurement-variables can be kept and investigated further by filtering the metric-results. The results can be visualized using ggplot2.
The general workflow and the way the different components of this package interact, can be seen in the introduction vignette.