Seaborn - grids and customization


In [4]:
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
iris= sns.load_dataset('iris')

sepal_length sepal_width petal_length petal_width species
0 5.1 3.5 1.4 0.2 setosa
1 4.9 3.0 1.4 0.2 setosa
2 4.7 3.2 1.3 0.2 setosa
3 4.6 3.1 1.5 0.2 setosa
4 5.0 3.6 1.4 0.2 setosa


Pairgrid is similar to pairplot, except, it returns back an empty grid that you can fill up with desired plots later. Refresher on pariplot

In [6]:
grd = sns.PairGrid(data=iris)

#then you can assign what you want plotted for diagonal, above diagonal, below diagonal.
# when mapping, pass just function pointers, dont call the function itself.

<seaborn.axisgrid.PairGrid at 0x11ea36f98>

lmplot() for scatter and regression per category

Sometimes, you need to do a joinplot() but split it by some categorical column. You can custom build it using FacetGrid shown in next section. However, seaborn provides a convenience function called lmplot(). Note: In previous pages, you created lmplots() for just 2 columns without any category.

In [330]:
pgrid = sns.lmplot(x='min_season', y='min_pressure_merged', 
                   col='any_basin', # the column by which you need to split - needs to be categorical
                   col_wrap=3, # number of columns per row
                   sharex=False, sharey=False, # will repeat ticks, coords for each plot
                   line_kws={'color':'green'} # symbol for regression line


During EDA, you want to find the distribution of data by sub-categories, sub-conditions. You can do so by building FacetGrids. As it means, you get a grid for every facet of the data.

In [7]:
#load tips data
tips = sns.load_dataset('tips')

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

Suppose we want to visualize total_bill by time of day and wheather or not it was a smoker. You need to filter data out then make dist plots. YOu can do all of that in 1 step with FacetGrids.

In [12]:
#for each unique value in `time` you get a row and 
# each unique value in `smoker` you get a col
fg = sns.FacetGrid(data=tips, row='time', col='smoker')

#now map a plot for each of the grid, 'total_bill')

<seaborn.axisgrid.FacetGrid at 0x11db6aa90>

Customizing grids

If you dont like the pale blue background of seaborn plots, you can modify that with set_style.

Note: Using set_style() will control the appearance for the entire notebook and all future plots

In [16]:
sns.set_style(style='ticks') #ticks, white, dark, darkgrid, whitegrid
#redraw the facet grid from above
fg = sns.FacetGrid(data=tips, row='time', col='smoker')

#now map a plot for each of the grid, 'total_bill')

<seaborn.axisgrid.FacetGrid at 0x11db832b0>

Fig and font size

You can use matplotlib figsize but have to specify that as a context as well.

In [24]:
plt.figure(figsize=(5,5)) #generate a fig, sns will piggyback this with the plot

<matplotlib.axes._subplots.AxesSubplot at 0x1204cd828>

Using seaborn context

You can use the set_context() to pick sizing templates

In [31]:
sns.set_context(context='poster', font_scale=0.8)
# valid contexts = paper, notebook, talk, poster - 
# with notebook being 1:1 and paper being smaller and poster being largest

#draw the facet grid
fg = sns.FacetGrid(data=tips, row='smoker', col='time'), 'total_bill')

<seaborn.axisgrid.FacetGrid at 0x120fccd68>

Another way to set the size is to access the fig handle direclty

In [35]:
#draw the facet grid
fg = sns.FacetGrid(data=tips, row='smoker', col='time')
#set the size
fg.fig.set_size_inches(w=10, h=10)
#plot the fig, 'total_bill')

<seaborn.axisgrid.FacetGrid at 0x122589ef0>