Pandas seaborn heatmap: text label horizontal

Question

Iam new to the seaborn library. I created a heatmap for cohort analysis. The text of y label is vertical. Instead I would love to show the text label horizontal. Thanks a lot for your support - Much appreciated. The red marked should be rotated, so the text is horizontal. I think this is not exactly y label.

**dataframe: df_cohort**
        cohort  order_month n_customers period_number
1   0   2023-01 2023-01 1018    0
2   1   2023-01 2023-02 384 1
3   2   2023-01 2023-03 502 2
4   3   2023-01 2023-04 423 3
5   4   2023-01 2023-05 468 4
6   5   2023-01 2023-06 371 5
7   6   2023-01 2023-07 332 6
8   7   2023-01 2023-08 226 7
9   8   2023-01 2023-09 351 8
10  9   2023-01 2023-10 321 9
11  10  2023-01 2023-11 262 10
12  11  2023-01 2023-12 221 11
13  12  2023-02 2023-02 621 0
14  13  2023-02 2023-03 179 1
15  14  2023-02 2023-04 205 2
16  15  2023-02 2023-05 235 3
17  16  2023-02 2023-06 184 4
18  17  2023-02 2023-07 142 5
19  18  2023-02 2023-08 84  6
20  19  2023-02 2023-09 166 7
21  20  2023-02 2023-10 157 8
22  21  2023-02 2023-11 102 9
23  22  2023-02 2023-12 92  10
24  23  2023-03 2023-03 412 0
25  24  2023-03 2023-04 111 1
26  25  2023-03 2023-05 124 2
27  26  2023-03 2023-06 112 3
28  27  2023-03 2023-07 72  4
29  28  2023-03 2023-08 42  5
30  29  2023-03 2023-09 88  6
31  30  2023-03 2023-10 69  7
32  31  2023-03 2023-11 56  8
33  32  2023-03 2023-12 43  9
34  33  2023-04 2023-04 211 0
35  34  2023-04 2023-05 48  1
36  35  2023-04 2023-06 50  2
37  36  2023-04 2023-07 43  3
38  37  2023-04 2023-08 25  4
39  38  2023-04 2023-09 40  5
40  39  2023-04 2023-10 35  6
41  40  2023-04 2023-11 26  7
42  41  2023-04 2023-12 16  8
43  42  2023-05 2023-05 171 0
44  43  2023-05 2023-06 41  1
45  44  2023-05 2023-07 32  2
46  45  2023-05 2023-08 23  3
47  46  2023-05 2023-09 37  4
48  47  2023-05 2023-10 35  5
49  48  2023-05 2023-11 24  6
50  49  2023-05 2023-12 14  7
51  50  2023-06 2023-06 139 0
52  51  2023-06 2023-07 23  1
53  52  2023-06 2023-08 18  2
54  53  2023-06 2023-09 34  3
55  54  2023-06 2023-10 24  4
56  55  2023-06 2023-11 33  5
57  56  2023-06 2023-12 16  6
58  57  2023-07 2023-07 80  0
59  58  2023-07 2023-08 16  1
60  59  2023-07 2023-09 18  2
61  60  2023-07 2023-10 14  3
62  61  2023-07 2023-11 9   4
63  62  2023-07 2023-12 16  5
64  63  2023-08 2023-08 31  0
65  64  2023-08 2023-09 8   1
66  65  2023-08 2023-10 12  2
67  66  2023-08 2023-11 8   3
68  67  2023-08 2023-12 6   4
69  68  2023-09 2023-09 84  0
70  69  2023-09 2023-10 16  1
71  70  2023-09 2023-11 20  2
72  71  2023-09 2023-12 13  3
73  72  2023-10 2023-10 60  0
74  73  2023-10 2023-11 14  1
75  74  2023-10 2023-12 9   2
76  75  2023-11 2023-11 35  0
77  76  2023-11 2023-12 3   1
78  77  2023-12 2023-12 43  0

cohort_pivot = df_cohort.pivot_table(index = 'cohort',
                                     columns = 'period_number',
                                     values = 'n_customers')

cohort_size = cohort_pivot.iloc[:,0]
retention_matrix = cohort_pivot.divide(cohort_size, axis = 0)

with sns.axes_style("white"):
    fig, ax = plt.subplots(1, 2, figsize=(12, 9), sharey=True, gridspec_kw={'width_ratios': [1, 11]})
    
    # retention matrix
    sns.heatmap(retention_matrix, 
                mask=retention_matrix.isnull(), 
                annot=True, 
                fmt='.0%', 
                cmap='RdYlGn', 
                ax=ax[1])
    ax[1].set_title('Monthly Cohorts: User Retention', fontsize=16)
    ax[1].set(xlabel='# of periods',
              ylabel='')

    # cohort size
    cohort_size_df = pd.DataFrame(cohort_size).rename(columns={0: 'cohort_size'})
    white_cmap = mcolors.ListedColormap(['white'])
    sns.heatmap(cohort_size_df, 
                annot=True, 
                cbar=False, 
                fmt='g', 
                cmap=white_cmap, 
                ax=ax[0])

    fig.tight_layout()

Answer 1

To rotate ticklabels, you'll need to rotate them manually or use the pyplot.setp helper function. You may also need to set their horizontal alignment parameter to 'right' to ensure they line up correctly after rotating.

import matplotlib.pyplot as plt

fig, ax = plt.subplots()

#...

# approach (recommended matplotlib)
ax.tick_params(axis='y', which='major', rotation=0)

# approach 1
for label in ax.get_yticklabels():
    label.set(rotation=0, ha='right')

# approach 2
plt.setp(ax.get_yticklabels(), rotation=0, ha='right')

Adding that into your code snippet with some randomly generated data produces:

import matplotlib.pyplot as plt
import matplotlib.colors as mcolors
from numpy.random import default_rng
import pandas as pd
import seaborn as sns

rng = default_rng(0)
retention_matrix = pd.DataFrame(rng.random(100).reshape(10, 10))
cohort_size = pd.Series(
    rng.integers(10, 100, size=len(retention_matrix)),
    index=pd.date_range('2000-01', freq='MS', periods=10).strftime('%Y-%m')
)


with sns.axes_style("white"):
    fig, ax = plt.subplots(1, 2, figsize=(12, 9), sharey=True, gridspec_kw={'width_ratios': [1, 11]})

    # retention matrix
    sns.heatmap(retention_matrix,
                mask=retention_matrix.isnull(),
                annot=True,
                fmt='.0%',
                cmap='RdYlGn',
                ax=ax[1])
    ax[1].set_title('Monthly Cohorts: User Retention', fontsize=16)
    ax[1].set(xlabel='# of periods',
              ylabel='')

    # cohort size
    cohort_size_df = pd.DataFrame(cohort_size).rename(columns={0: 'cohort_size'})
    white_cmap = mcolors.ListedColormap(['white'])
    sns.heatmap(cohort_size_df,
                annot=True,
                cbar=False,
                fmt='g',
                cmap=white_cmap,
                ax=ax[0])
    ax[0].tick_params(axis='y', which='major', rotation=0)

    fig.tight_layout()
    plt.show()