CODEWITHSUNDEEP
pythonnumpymatplotlibseaborn
mvd
mvd

Reputation: 1199

customizing facetgrid xticks and labels in seaborn

how can the properties of ticklabels be customized inside a FacetGrid in seaborn? in the example below i wanted to change the fontsize of only the x-axis ticklabels:

import matplotlib.pylab as plt
import seaborn as sns
df1 = pandas.DataFrame({"x": np.random.rand(100),
                        "y": np.random.rand(100)})
df1["row"] = "A"
df1["col"] = "1"
df2 = pandas.DataFrame({"x": np.random.rand(100),
                       "y": np.random.rand(100) + 1})
df2["row"] = "A"
df2["col"] = "2"
df = pandas.concat([df1, df2])
g = sns.FacetGrid(df, row="row", col="col", hue="col")
g.map(plt.scatter, "x", "y")
print g.axes
row, col = g.axes.shape
for i in range(row):
    for j in range(col):
        ax = g.axes[i, j]
        print "ax: ", ax.get_xticklabels()
        ax.set_xticklabels(ax.get_xticklabels(), fontsize=6)

my solution iterates through the axes array of the FacetGrid object which seems unnecessarily complicated. i know fontscale can be set globally but i only want to change the x-axis ticklabel fonts. is there a better way to do this? and is there also a way to customize the style of labels to be scientific notation?

Upvotes: 1

Views: 5008

Answers (1)

paulwasit
paulwasit

Reputation: 426

You can simplify the loop by using the flatten method. I have also added a way to format in scientific notation. .0e means scientific notation with 0 decimals.

import matplotlib.ticker as tkr

for ax in g.axes.flatten():
    ax.set_xticklabels(ax.get_xticklabels(), fontsize=6)
    ax.xaxis.set_major_formatter(
        tkr.FuncFormatter(lambda x, p: "{:.0e}".format(x)))

Upvotes: 4

Related Questions