Adjust y-axis in Seaborn multiplot

Question

I'm plotting a CSV file from my simulation results. The plot has three graphs in the same figure fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(24, 6)).

However, for comparison purposes I want the y-axis in all graphs starting at zero and the ending at a specific value. I tried the solution mentioned here from the Seaborn author. I don't get any errors, but the solution also does not work for me.

Here's my script:

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

fname = 'results/filename.csv'

def plot_file():
    fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(24, 6))
    df = pd.read_csv(fname, sep='\t')
    profits = \
        df.groupby(['providerId', 'periods'], as_index=False)['profits'].sum()

    # y-axis needs to start at zero and end at 10
    g = sns.lineplot(x='periods',
                     y='profits',
                     data=profits,
                     hue='providerId',
                     legend='full',
                     ax=axes[0])

    # y-axis need to start at zero and end at one
    g = sns.scatterplot(x='periods',
                        y='price',
                        hue='providerId',
                        style='providerId',
                        data=df,
                        legend=False,
                        ax=axes[1])
    # y-axis need to start at zero and end at one
    g = sns.scatterplot(x='periods',
                        y='quality',
                        hue='providerId',
                        style='providerId',
                        data=df,
                        legend=False,
                        ax=axes[2])

    g.set(ylim=(0, None))
    plt.show()

    print(g) # -> AxesSubplot(0.672059,0.11;0.227941x0.77)

The resulting figure is as follows:

How can I adjust each individual plot?

Answer 1

Based on the way you've written your code, you can refer to each subplot axis with g.axis and use g.axis.set_ylim(low,high). (A difference compared to the linked answer is that your graphs are not being plotted on a seaborn FacetGrid.)

An example using dummy data and different axis ranges to illustrate:

df = pd.DataFrame(np.random.uniform(0,10,(100,2)), columns=['a','b'])

fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(8,4))


g = sns.lineplot(x='a',
                 y='b',
                 data=df.sample(10),
                 ax=axes[0])
g.axes.set_ylim(0,25)

g = sns.scatterplot(x='a',
                    y='b',
                    data=df.sample(10),
                    ax=axes[1])
g.axes.set_ylim(0,3.5)

g = sns.scatterplot(x='a',
                    y='b',
                    data=df.sample(10),
                    ax=axes[2])
g.axes.set_ylim(0,0.3)

plt.tight_layout()
plt.show()