Painlessly visualize results of machine learning analyses

• Free software: BSD license

• TODO

Highly configurable scatterplot, allowing you to specify the hue, marker, shape, size, alpha, linewidth and edgecolor of a plot in a single command, similar to ggplot2 in R.

Instead of this:

tedious matplotlib code

You can do this:

import cupcake as cup

simple cupcake code

Combine cupcake with the statistical plotting library seaborn to create grids of configurable scatterplots.

import seaborn as sns
import cupcake as cup
sns.set(style='ticks', context='talk')

iris = sns.load_dataset('iris')

g = sns.FacetGrid(iris)
g.map_dataframe(cup.scatterplot, x='sepal_length', y='sepal_width',
                hue='species', alpha='petal_length', size='petal_width')

For all your dimensionality reduction needs! Given any high-dimensional dataset, perform dimensionality reduction and plot the result.