Closed willeppy closed 1 year ago
One potential design is for all columns in a dataframe you can use a dropdown, or dragging, to combine with other columns. Then this plot reactively updates so long as the input columns are still in the dataframe. This specification would happen entirely through the UI
Another option is to do so programatically:
import digautoprofiler
digautoprofiler.plot(df['x'], df['y'])
This seems worse than the first option at first glance
Putting here for tracking, but I think out of scope for now is considering 3D data. For instance, I have, x, y, z data that I want to plot together. I think for initial version we will limit to just bivariate
Want ability to plot multiple columns together rather than only univariate information. This is very common task so would see lots of use, however needs careful design so as not to violate the minimal nature of AutoProfiler currently.
Example
If I have dataframe with columns:
[x, y]
. I want to be able to plotx
vsy