search

datamole-ai / edvart

An open-source Python library for Data Scientists & Data Analysts designed to simplify the exploratory data analysis process. Using Edvart, you can explore data sets and generate reports with minimal coding.
https://datamole-ai.github.io/edvart/
MIT License
48 stars 7 forks source link

feat: treat categorical color column which is also numeric as categorical in parallel coordinates & parallel categories #191

Closed mbelak-dtml closed 11 months ago

mbelak-dtml commented 11 months ago

Previously, categorical color columns (e.g. with values {1,2,3}, such as edvart.example_datasets.dataset_auto()["origin"]) would be treated as numeric in multivariate analysis parallel coordinates & parallel categories sections, i.e. a continuous color scale would be used.

mbelak-dtml commented 11 months ago

The same issue is in pairplot. I would either modify the commit message so that it mentions that it changes the behaviour only for parallel categories/coordinates or change it everytwhere.

image

The issue is not there -- cylinders is numeric, but treated as categorical. Otherwise there would a continuous color scale, but there is a discrete color scale, which is correct.