An evolution of 'reshape2'. It's designed specifically for data
tidying (not general reshaping or aggregating) and works well with
'dplyr' data pipelines.