Data Visualization

import pandas as pd
import numpy as np
import statsmodels.api as sm
import seaborn as sns
import matplotlib.pyplot as plt

df = sm.datasets.get_rdataset("mtcars", "datasets", cache = True).data

df.head()

	mpg	cyl	disp	hp	drat	wt	qsec	vs	am	gear	carb
Mazda RX4	21.0	6	160.0	110	3.90	2.620	16.46	0	1	4	4
Mazda RX4 Wag	21.0	6	160.0	110	3.90	2.875	17.02	0	1	4	4
Datsun 710	22.8	4	108.0	93	3.85	2.320	18.61	1	1	4	1
Hornet 4 Drive	21.4	6	258.0	110	3.08	3.215	19.44	1	0	3	1
Hornet Sportabout	18.7	8	360.0	175	3.15	3.440	17.02	0	0	3	2

Seaborn seems to be the most efficient way to get decent looking exploratory plots in a hurry.

Line Plot

sns.lineplot(df, x = "mpg", y = "disp")

<AxesSubplot: xlabel='mpg', ylabel='disp'>

Line Plot by factor

Use the hue argument to break out factors into separate lines.

sns.lineplot(df, x = "mpg", y = "disp", hue = "am")

<AxesSubplot: xlabel='mpg', ylabel='disp'>

Line plot with linear trend

Mean/CI are automatic if you melt the df.

df_long = pd.melt(df, id_vars = "cyl", value_vars = "mpg")

df_long

	cyl	variable	value
0	6	mpg	21.0
1	6	mpg	21.0
2	4	mpg	22.8
3	6	mpg	21.4
4	8	mpg	18.7
5	6	mpg	18.1
6	8	mpg	14.3
7	4	mpg	24.4
8	4	mpg	22.8
9	6	mpg	19.2
10	6	mpg	17.8
11	8	mpg	16.4
12	8	mpg	17.3
13	8	mpg	15.2
14	8	mpg	10.4
15	8	mpg	10.4
16	8	mpg	14.7
17	4	mpg	32.4
18	4	mpg	30.4
19	4	mpg	33.9
20	4	mpg	21.5
21	8	mpg	15.5
22	8	mpg	15.2
23	8	mpg	13.3
24	8	mpg	19.2
25	4	mpg	27.3
26	4	mpg	26.0
27	4	mpg	30.4
28	8	mpg	15.8
29	6	mpg	19.7
30	8	mpg	15.0
31	4	mpg	21.4

sns.lineplot(df_long, x = "cyl", y = "value")

<AxesSubplot: xlabel='cyl', ylabel='value'>

Or, use lmplot to git a linear model like you’d get with geom_smooth(method = lm).

sns.lmplot(df, x = "mpg", y = "disp")

<seaborn.axisgrid.FacetGrid at 0x7f5d567270d0>

Box Plot

sns.boxplot(df, x = "cyl", y = "mpg")

<AxesSubplot: xlabel='cyl', ylabel='mpg'>

Histogram

sns.histplot(df, x = "wt")

<AxesSubplot: xlabel='wt', ylabel='Count'>

Density Plot

sns.kdeplot(df, x = "disp")

<AxesSubplot: xlabel='disp', ylabel='Density'>

Heatmap

sns.heatmap(df[["disp", "hp"]])

<AxesSubplot: >

Multiple Variable Plots

KDE

sns.kdeplot(df.loc[:, ["mpg", "wt"]])

<AxesSubplot: ylabel='Density'>

Lineplot

sns.lineplot(df.loc[:, ["mpg", "wt"]])

<AxesSubplot: >

Faceting

# create three empty spots
grid = sns.FacetGrid(data = df, col = "cyl", col_wrap=2)

# puts a historgram on each of them
grid.map(sns.histplot, "wt")

<seaborn.axisgrid.FacetGrid at 0x7f5d55cf2590>

The initial display is automatic. If you want to show the same plot again, access the figure property of the object.

# just typing it out gives the object metadata
grid

<seaborn.axisgrid.FacetGrid at 0x7f5d55cf2590>

grid.figure

Tweaking Plots

Axis Labels

The plot we made of weight and mpg had mostly unusable x tick labels. Let’s revist it.

p_line = sns.lineplot(df.loc[:, ["mpg", "wt"]])
p_line.figure

The syntax is a little awkward. Essentially there is a set method, and you use a get method to retrieve the labels to pass into it, specifying a rotation.

# set what you get from the get method v--here
p_line.set_xticklabels(p_line.get_xticklabels(), rotation = 45)
p_line.figure

/tmp/ipykernel_9407/667538073.py:2: UserWarning: FixedFormatter should only be used together with FixedLocator
  p_line.set_xticklabels(p_line.get_xticklabels(), rotation = 45)

They still conflict a little. We can make them a little smaller overall. The technique is the same, just setting a different property.

p_line.set_xticklabels(p_line.get_xticklabels(), size = 5)
p_line.figure

/tmp/ipykernel_9407/3729791072.py:1: UserWarning: FixedFormatter should only be used together with FixedLocator
  p_line.set_xticklabels(p_line.get_xticklabels(), size = 5)

Title

p_line.set(title = "0_o")

p_line.figure

Color Schemes

Discrete

Seaborn lets you preview color palettes by calling them as a function argument to sns.color_palette.

sns.color_palette("dark")

The plotting functions will then have arguments for color scheming:

p_box = sns.boxplot(df, x = "cyl", y = "mpg", palette = "dark")

Continuous

sns.color_palette("mako", as_cmap = True)

mako

under

bad

over

sns.heatmap(df[["disp", "hp"]], cmap = "mako")

<AxesSubplot: >