I've found that the trend detection doesn't work when value_col is between [0, 1]. Scaling value_col to a larger value works fine. Given normalised datasets are relatively likely, this seems like a fairly common use case.
Environment:
Python: 3.14.3 (Anaconda)
OS: Macos 26.3.1
Library version: 1.1.10
I've found that the trend detection doesn't work when
value_colis between [0, 1]. Scalingvalue_colto a larger value works fine. Given normalised datasets are relatively likely, this seems like a fairly common use case.Environment:
Python: 3.14.3 (Anaconda)
OS: Macos 26.3.1
Library version: 1.1.10