Seaborn¶

Searborn is a Python library for Statistical Data Visualization. It provides a high-level interface and many "out-of-the-box" plotting functionality for easy exploration. Seaborn.jl is a Julia wraper of the python library.

Imports¶

In [39]:

using Seaborn
using Pandas
using PyPlot
using PyCall

@pyimport numpy

Visualizing linear relationships:¶

Seaborn is not a regression library itself. For quantitative measures related to the fit of regression models, you should use GLM.jl. However, Seaborn provides regression plots in seaborn that helps emphasizing patterns in a dataset during exploratory data analyses.

Functions for linear-regreassion models¶

regplot: In the simplest invocation, draw a scatterplot of two variables, x and y, and then fit the regression model y ~ x and plot the resulting regression line and a 95% confidence interval for that regression. Inputs x,y can be in a variaerty of formats.
lmplot: Uses regplot.

Inputs must be Pandas.DataFrame format.

jointplot: Use regplot together with dictribution plots to provide and alternative visualization of the relationship

More info

In [40]:

#for some strange reason loading doesn't work the first time
tips = nothing
try
    tips = load_dataset("tips");
catch
    tips = load_dataset("tips");
end


head(tips)

Out[40]:

   day     sex  size smoker    time   tip  total_bill
0  Sun  Female     2     No  Dinner  1.01       16.99
1  Sun    Male     3     No  Dinner  1.66       10.34
2  Sun    Male     3     No  Dinner  3.50       21.01
3  Sun    Male     2     No  Dinner  3.31       23.68
4  Sun  Female     4     No  Dinner  3.61       24.59

In [43]:

g = regplot(x="total_bill", y="tip", data=tips)
title("Total Bill vs. Tip") #current active figure PyPlot
# alternatively
# g[:figure][:axes][1][:set_title]("Total Bill vs. Tip")

Out[43]:

PyObject <matplotlib.text.Text object at 0x32307d3c8>

One discrete variable:¶

Simple scatter plot is not optimal
Add random noise "jitter" to the points (only on plot) to see their distribution more clearely
Collapse the observations to the mean along with a confidence interval

In [4]:

f, (ax1, ax2,ax3)  = subplots(1, 3, sharey=true)
regplot(x="size", y="tip", data=tips, ax=ax1);
regplot(x="size", y="tip", data=tips, x_jitter=.05, ax=ax2);
regplot(x="size", y="tip", data=tips, x_estimator=numpy.mean, ax=ax3);

Higher order models¶

In [5]:

anscombe = load_dataset("anscombe");
head(anscombe)

lmplot(x="x", y="y", data=query(anscombe, "dataset == 'II'"),
           ci=nothing, scatter_kws=Dict("s"=> 80));

In [6]:

lmplot(x="x", y="y", data=query(anscombe, "dataset == 'II'"),
           ci=nothing, scatter_kws=Dict("s"=> 80), order =2);

Joint Plot with marginal distributions¶

In [44]:

jointplot(x="total_bill", y="tip", data=tips, kind="reg")
savefig("./test.svg")

Conditioning on another variable¶

Must use lmplot instead of regplot

In [8]:

lmplot(x="total_bill", y="tip", hue="smoker", data=tips);

In [9]:

lmplot(x="total_bill", y="tip", col="day", data=tips,
           aspect=.5);