An informative RDI categorical distribution plot inspired by Violin, Bean and Pirate Plots.
(RDI = Raw data + Descriptive statistics + Inferential statistics)
-
Like Violin plots, it shows smoothed kernel density curves, revealing information which would be hidden in boxplots, for example presence of multiple "peaks" ("modes") in the distribution "mountain".
-
Like Bean plots, it shows the raw data, drawn as a cloud of points. By default all data points are shown but you can optionally control this and limit the display to a subset of the data.
-
Like Pirate plots, it marks confidence intervals (either from Student's T or as Bayesian Highest Density Intervals or as interquantile ranges) for the probable position of the true population mean.
Since by default it does not symmetrically mirror the density curves, it allows immediate comparisions of distributions side-by-side.
cmplot is pure python code. It has no platform-specific dependencies and should thus work on all platforms. It requires the packages plotly numpy scipy pandas. The latest version of cmplot can be installed by typing either:
pip3 install cmplot(from Python Package Index)
or:
pip3 install git+git://github.com/g-insana/cmplot.py.git
(from GitHub).
There is also a version in Julia.
>>> import plotly.graph_objects as go
>>> from cmplot import cmplot
#call the cmplot directly inside a plotly Figure function as:
>>> go.Figure(*cmplot(mydataframe,xcol="xsymbol"))
#alternatively get traces and layout as separate variables, so that you can modify them or combine with others before passing them to Figure() function:
>>> (traces,layout)=(cmplot(mydataframe,xcol="xsymbol"))
#[...] do something with traces/layout
>>> go.Figure(traces,layout) #plot itcmplot is licensed under the GNU Affero General Public License.
(c) Copyright Giuseppe Insana, 2019-