Pandas Dataframe examples: Plotting Histograms

Pandas Dataframe examples: Plotting Histograms

Last updated:
Table of Contents

All code available on this jupyter notebook

Histogram of column values

You can also use numpy arange to create bins automatically: np.arange(<start>,<stop>,<step>)

Example: Plot histogram of values in column "age"

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({
    'name':['john','mary','peter','jeff','bill','lisa','jose'],
    'age':[23,78,22,19,45,33,20],
    'gender':['M','F','M','M','M','F','M'],
    'state':['california','dc','california','dc','california','texas','texas'],
    'num_children':[2,0,0,3,2,1,4],
    'num_pets':[5,1,0,5,2,2,3]
})

df[['age']].plot(kind='hist',bins=[0,20,40,60,80,100],rwidth=0.8)
plt.show()

source-dataframe Source dataframe
age-by-bins The most common age group is between 20 and 40 years old

Relative histogram (density plot)

In other words, plot the density of the values, regardless of scale.

This is useful to get a relative histogram of values, which allows you to compare distributions of different sizes and scales.

Just pass density=True to df.plot(kind='hist'):

import scipy.stats as st
from matplotlib import pyplot as plt

# standard normal distribution
dist_1 = st.norm(loc=0.0,scale=1.0)

df1 = pd.DataFrame({
    'value_1': dist_1.rvs(10000),
})

df1.plot(kind='hist', density=True)

pandas-histogram-no-density If you don't pass density=True
the y-axis is just the absolute frequency of values
  
pandas-histogram-density-plot By passing density=True the plot is now
a density plot, and the y-axis now represents relative frequency.

Two plots one the same Axes

We frequently want to plot two (or more) distributions together so we can compare them even if the samples have different sizes

Pass the same ax to both and use alpha to make the plots transparent:

import scipy.stats as st
from matplotlib import pyplot as plt

# just two dummy distributions
dist_1 = st.norm(loc=0.0,scale=5.0)
dist_2 = st.norm(loc=5,scale=1.0)

nums_1 = dist_1.rvs(1000)
nums_2 = dist_2.rvs(50000) # note that it's a much larger sample

df1 = pd.DataFrame({'value_1': nums_1})

df2 = pd.DataFrame({'value_2': nums_2})

ax = plt.gca()

# note we're using density=True because the two samples
# have different sizes
df1.plot(kind='hist', ax=ax, density=True, alpha=0.5)
df2.plot(kind='hist', ax=ax, density=True, alpha=0.5)

two-distributions-on-the-same-chart-pandas-matplotlib By plotting both distributions on the same plot and
by setting alpha you can easily compare both and
see where they overlap (around x=5)

Group large values

In other words, truncate large values after a given point into a single bucket, called "+∞":

Example: group values larger than 100 into a separate column:

sample-data sample data used: note that there are rows where age is larger than 100!

import pandas as pd
import random
import matplotlib.pyplot as plt
import numpy as np

# install via pip install faker
from faker import Faker
fake = Faker()

df = pd.DataFrame([{
    'name':fake.name(),
    'age': random.randint(1,120), # generate age from 1 to 120
} for i in range(0,100)]) # dataframe should have 100 rows

# define the bins
step_size          = 20
max_tick           = 100 
original_bins      = np.arange(0, max_tick+1, step_size)
new_bins           = np.append(original_bins, [max_tick+step_size+1])
max_value          = max(original_bins)

# function to format the label
def format_text(current_value, max_value, label):
    return label if current_value > max_value else current_value

df[['age']].plot(kind='hist', bins=new_bins, rwidth=0.8)

current_ticklabels = plt.gca().get_xticks()

plt.gca().set_xticklabels([format_text(x, max_value, '+∞') for x in current_ticklabels])

plt.show()

naivelly-plotting If you naively plot histogram with 0-20 buckets,
some of the data will be missing!
  
plotting-after-truncating-values Values larger than 100 grouped into the new bucket