Plotting a dataframe as both a 'hist' and 'kde' on the same plot

Question

I have a pandas dataframe with user information. I would like to plot the age of users as both a kind='kde' and on kind='hist' on the same plot. At the moment I am able to have the two separate plots. The dataframe resembles:

member_df=    
user_id    Age
1          23
2          34
3          63 
4          18
5          53  
...

using

ax1 = plt.subplot2grid((2,3), (0,0))
member_df.Age.plot(kind='kde', xlim=[16, 100])
ax1.set_xlabel('Age')

ax2 = plt.subplot2grid((2,3), (0,1))
member_df.Age.plot(kind='hist', bins=40)
ax2.set_xlabel('Age')

ax3 = ...

I understand that the kind='kde' will give me frequencies for the y-axis whereas kind='kde' will give a cumulative distribution, but is there a way to combine both and have the y-axis be represented by the frequencies?

Solution 1

pd.DataFrame.plot() returns the ax it is plotting to. You can reuse this for other plots.

Try:

ax = member_df.Age.plot(kind='kde')
member_df.Age.plot(kind='hist', bins=40, ax=ax)
ax.set_xlabel('Age')

example
I plot hist first to put in background
Also, I put kde on secondary_y axis

import pandas as pd
import numpy as np


np.random.seed([3,1415])
df = pd.DataFrame(np.random.randn(100, 2), columns=list('ab'))

ax = df.a.plot(kind='hist')
df.a.plot(kind='kde', ax=ax, secondary_y=True)

---

response to comment
using subplot2grid. just reuse ax1

import pandas as pd
import numpy as np

ax1 = plt.subplot2grid((2,3), (0,0))

np.random.seed([3,1415])
df = pd.DataFrame(np.random.randn(100, 2), columns=list('ab'))

df.a.plot(kind='hist', ax=ax1)
df.a.plot(kind='kde', ax=ax1, secondary_y=True)

Solution 2

In case you want it for all the columns of your dataframe:

fig, ax = plt.subplots(8,3, figsize=(20, 50)) 
# you can change the distribution, I had 22 columns, so 8x3 is fine to me
fig.subplots_adjust(hspace = .2, wspace=.2, )

ax = ax.ravel()

for i in range(len(I_df.columns)):
    ax[i] = I_df.iloc[:,i].plot(kind='hist', ax=ax[i])
    ax[i] = I_df.iloc[:,i].plot(kind='kde', ax=ax[i], secondary_y=True)
    plt.title(I_df.columns[i])

I hope it helps :)

Solution 3

It is better and even simpler to use seaborn.displot. Prior proposed solutions had KDE plot appear a little "shifted up" for me. seaborn.distplot accurately lined up zeros between hist and kde plots.
import seaborn as sns sns.displot(df.a)

Comments

[Lukasz]

I've tested the code and attempted to modify it slightly for what I need. This does the trick for when I only have those two plots to consider. When I attempt to include into sublot2grid it fails to produce the same outcome, it is only reproduces the histogram.

[piRSquared]

@Lukasz you want to use the same ax in any case.