splineplot

A tiny Python package for plotting splines

splineplot is a tiny Python package that uses statsmodels to fit a spline to 1-dimensional data and plot it, along with a scatter plot. It can be used on its own, or in conjunction with Seaborn; its interface is similar to that of regplot.

import seaborn as sns

iris = sns.load_dataset("iris")

g = sns.FacetGrid(
    iris,
    hue="species",
    aspect=1.3,
    height=4,
)

g.map(
    splineplot,
    "sepal_length",
    "sepal_width",
    alpha=0,
    scatter_kws={"s": 3},
)
_ = g.add_legend()

A Note On alpha

The alpha argument to splineplot determines the smoothing penalty used when the spline is fit to the y data. It's important to consider overfitting when setting this value; in general a value of 0 with modest dataset sizes is almost guaranteed to overfit.