CODEWITHSUNDEEP
pythonseaborn
dfahsjdahfsudaf
dfahsjdahfsudaf

Reputation: 481

Hue on Seaborn Barplot

I am looking to plot 'surgeon id' on the x-axis and 'Total Cases' on the y-axis. Within each bar, I would like to color the bar in relation to the % for Treatment X and Y.

treatment  surgeon_id initial_severity  Total Cases  Treatment X%  Treatment Y%
    0                   2            MINOR         16.0      6.250000     93.750000
    1                   3            MINOR         41.0     36.585366     63.414634
    2                   4            MINOR         69.0     33.333333     66.666667
    3                   5            MINOR        102.0     34.313725     65.686275
    4                   6            MINOR        348.0     11.781609     88.218391
    5                   7            MINOR         20.0     40.000000     60.000000
    6                   8            MINOR         90.0     28.888889     71.111111
    7                   9            MINOR         41.0     17.073171     82.926829
    8                  10            MINOR        139.0     24.460432     75.539568
    9                  11            MINOR         46.0     23.913043     76.086957
    10                 12            MINOR         35.0     20.000000     80.000000
    11                 13            MINOR         96.0     29.166667     70.833333
    12                 16            MINOR         51.0     21.568627     78.431373
    13                 17            MINOR         90.0     27.777778     72.222222
    14                 18            MINOR        100.0     24.000000     76.000000
    15                 19            MINOR         78.0     28.205128     71.794872
    16                 20            MINOR         96.0     25.000000     75.000000
    17                 21            MINOR         23.0     26.086957     73.913043
    18                 22            MINOR          9.0     33.333333     66.666667
    19                 25            MINOR         78.0     30.769231     69.230769
    20                 28            MINOR          7.0     14.285714     85.714286
    21                 30            MINOR         70.0     25.714286     74.285714
    22                 31            MINOR         21.0     42.857143     57.142857
    23                 32            MINOR        101.0     29.702970     70.297030
    24                 41            MINOR          1.0    100.000000      0.000000
    25                 43            MINOR         15.0      0.000000    100.000000

Current plotting code:

sns.barplot(x="surgeon_id", y="Total Cases", data=minor_treatment_choice)

How can I color each bar in relation to the % for Treatment X and Y? I would also like the color for Treatment X and Y within the bars to be the same color for every surgeon_id.

Thanks!

Upvotes: 0

Views: 7916

Answers (1)

JohanC
JohanC

Reputation: 80279

You can add two extra columns, where you multiply the percentages with the total cases.

df["Treatment X"] = df["Treatment X%"] * df["Total Cases"] / 100
df["Treatment Y"] = df["Treatment Y%"] * df["Total Cases"] / 100

Then, with pandas you could create a stacked bar plot:

df.set_index("surgeon_id")[["Treatment X", "Treatment Y"]].plot.bar(color=['crimson', 'dodgerblue'], stacked=True)

To create stacked bars with Seaborn, you can first draw the lower bars, and in a second call add the upper bars setting their bottom to the top of the lower bars:

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

df["Treatment X"] = df["Treatment X%"] * df["Total Cases"] / 100
df["Treatment Y"] = df["Treatment Y%"] * df["Total Cases"] / 100

ax = sns.barplot(data=df, x="surgeon_id", y=df["Treatment X"], color='crimson', label="Treatment X")
sns.barplot(data=df, x="surgeon_id", y=df["Treatment Y"], bottom=df["Treatment X"], color='dodgerblue',
            label="Treatment Y", ax=ax)
ax.legend()
plt.show()

resulting plot

Upvotes: 2

Related Questions