CODEWITHSUNDEEP
pythonplotseaborn
00__00__00
00__00__00

Reputation: 5327

highlight part of a jointplot in seaborn

I am generating some professionally looking figures in seaborn.

For instance:

import seaborn as sns
sns.set_style('white')
sns.set_context('paper', font_scale=2)
sns.set_style('ticks', {'axes.edgecolor': '0',  
                        'xtick.color': '0',
                        'ytick.color': '0'})

g = sns.jointplot(x=np.random.rand(1000,3), y=np.random.rand(1000,3),
                   marginal_kws=dict(bins=50, rug=True,color="k"),
                   annot_kws=dict(stat="r"),
                   s=40, edgecolor="k",color='w', linewidth=1)
sns.despine(offset=5, trim=True)
g.fig.set_figheight(12)
g.fig.set_figwidth(12)

gives as output: enter image description here

Now, I would like to highlight part of the figure.

I would like the top histogram to have different colors for a certain range e.g. all bars where x<0.25 should be red, and corresponding point in the scatter plot as well.

How to obtain that?

I could repeat the plot operation using a subset of the original data, but then the histogram scaling would different.

Upvotes: 0

Views: 292

Answers (1)

SSharma
SSharma

Reputation: 953

Note: This is not a complete answer
For scatter points you can annotate the points which are less than a given value:

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

df = pd.DataFrame(columns=("x", "y", "x_cor_to_ann"))
df['x'] = pd.Series(np.random.randn(10))
df['y'] = pd.Series(np.random.randn(10))
df['x_cor_to_ann'] = df['x'] <= 0.25
print(df)

g = sns.jointplot("x", "y", data=df, kind="reg",
                  marginal_kws=dict(bins=50, rug=True,color="k"),
                  annot_kws=dict(stat="r"),
                   color='b')

highlight = df[df['x']<=0.25]

def annotate(data_row):
    r = data_row[1]
    plt.gca().annotate(data_row[0], xy=(r["x"], r["y"]), 
            xytext=(2,2) , textcoords ="offset points")

for data_row in highlight.iterrows():
    annotate(data_row)

plt.show()
sns.despine(offset=5, trim=True)
g.fig.set_figheight(12)
g.fig.set_figwidth(12)

So if your dataframe looked like this:

      x         y         x_cor_to_ann
0  1.180725  0.368682         False
1  1.816001  0.104237         False
2  1.074462  1.253019         False
3 -0.219031 -0.670320          True
4  0.402527 -0.899265         False
5 -0.960773  2.942559          True
6  0.067588  0.508284          True
7 -0.051982 -0.194170          True
8 -0.016429  0.101746          True
9  1.607429  0.736233         False

Which should look like this: enter image description here
Which should annotate the rows where x<=0.25 and the rest are not annotated. You can of course do more with annotation rather than annotating the index number.
With the histogram, I'm not sure how to do that, sorry :(. The seaborn doc doesn't help either.

Upvotes: 1

Related Questions