Splitting matplotlib bar graph into 3

Question

Plotting the severity of accident by the month which it occurred. My data has 3 values for severity (0,1,2) all which are under one graph. I want to create three separate graphs for each severity value.

month = df.groupby(['Month','Severity']).size().unstack()
print(month)
month.plot(kind='bar')
plt.legend(title = 'Severity')
plt.show()

Answer 1

Here's my solution using Matplotlib:

# Import packages
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Create random data - Just convert your data to a Pandas dataframe
months = np.array(['January', 'February', 'March', 'April', 'May', 'June', 'July',
                      'August', 'September', 'October', 'November', 'December'])
numMonths = months.shape[0]

sev0 = np.random.randint(13000, 15500, (numMonths,))
sev1 = np.random.randint(250, 350, (numMonths,))
sev2 = np.random.randint(20, 35, (numMonths,))

d = {'Month': months, 'Sev_0': sev0, 'Sev_1': sev1, 'Sev_2': sev2}
df = pd.DataFrame(data=d)

# Create plots
for i in range(3):
    yData = "Sev_" + str(i)
    plt.figure(figsize=(12,10))
    
    plt.title(yData)
    plt.bar(df.Month, df[yData].to_numpy())
    plt.xlabel('Month')
    plt.ylabel('Number of Accidents')

where df is your data (replace my random data generation with your actual data).

Here is a sample output:

If you want to get fancy, you can use Plotly, which has hover and zoom features
Also, you can control which severity value is being plotted when all in one interactive graph.

import plotly
import plotly.graph_objs as go
from plotly.offline import init_notebook_mode, plot, iplot, download_plotlyjs
init_notebook_mode(connected=True)
plotly.offline.init_notebook_mode(connected=True)

dataAllSev = []
for i in range(3):
    yData = "Sev_" + str(i)
    
    dataAllSev.append(go.Bar(x=df.Month, y=df[yData].to_numpy(), name=yData))

fig = go.Figure(data=dataAllSev)
fig.update_layout(title="Number of Accidents by Severity Level", xaxis_title='Month', yaxis_title='Number of Accidents')
fig.show()

Here is the combined graph in Plotly: