You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
First of all, thanks for the work on this great package. We've just discovered it recently and it will surely provide some benefit for our daily work in our team.
Painpoint
Having numerical columns with a dense area of values and some extreme outliers, the default histogram does not provide any useful insight because there is basically only one bin with almost all values (dense area) and the other bins are almost empty (outliers).
Solution
Instead of fixed bin widths, it would be useful to use algorithms with variable bin widths to account for unevenly distributed densities. AstroML already has two implementations for this (see here). The algorithms are not to complicated, so that they could be vendored without introducing a new dependency (see here) @jakevdp.
The text was updated successfully, but these errors were encountered:
- Added Variable bin sizing via Bayesian Boxing (feature request [#216])
- PyCharm integration, console attempts to detect file type.
- Fixed bug [#215].
- Updated the `missingno` package to 0.4.2, fixing the font size in the `bar` diagram.
- Various optimizations
- Added Variable bin sizing via Bayesian Boxing (feature request [ydataai#216])
- PyCharm integration, console attempts to detect file type.
- Fixed bug [ydataai#215].
- Updated the `missingno` package to 0.4.2, fixing the font size in the `bar` diagram.
- Various optimizations
First of all, thanks for the work on this great package. We've just discovered it recently and it will surely provide some benefit for our daily work in our team.
Painpoint
Having numerical columns with a dense area of values and some extreme outliers, the default histogram does not provide any useful insight because there is basically only one bin with almost all values (dense area) and the other bins are almost empty (outliers).
Solution
Instead of fixed bin widths, it would be useful to use algorithms with variable bin widths to account for unevenly distributed densities. AstroML already has two implementations for this (see here). The algorithms are not to complicated, so that they could be vendored without introducing a new dependency (see here) @jakevdp.
The text was updated successfully, but these errors were encountered: