orange3
orange3 copied to clipboard
biolab
Group By: Add quartile outputs
What's your use case?
I would like to summarize large datasets through the 5-number set (minimum, Q1, median, Q3, maximum) + mean, but quartiles are not available in the Group By GUI.
What's your proposed solution?
Integrate quartile computations and 1 or 2 check boxes ("Quartiles" or "First quartile" / "Third quartile") in the GUI of the Group By widget.
Are there any alternative solutions?
These can be computed with a custom python script using pandas in the meantime.
I think this could be added to either Box Plot or Feature Statistics (which already outputs Statistics) instead. Box Plot could indeed have a Statistics output, since the widget already computes all the necessary information.
Indeed, it might be a good addition! Still, in my use case, I need per-class statistics, and I thought it would make sense in this widget, as it already computes multiple distribution statistics
It turns out Box Plot computes these statistics on the fly, so that's not really a good option. Feature Statistics might work, but yeah, I guess adding it to Group By cannot hurt.
Decision: We'll do it. The first column will contain Mean, Median, Q1, Q3, Min, Max.
