同一matplotlib图表上的负值栏

Question

I am trying to display 3 bar charts on the same plot. There is an issue with bars that have negative values though, because they are hanging down either from the top or from nowhere. Any ideas how to make it look nicer?

import pandas as pd
import matplotlib.pyplot as plt

x = range(6)
a1 = [-1, -4, -3, -6, 2, 8]
a2 = [ 4, 12, 8, 1, 10, 9]
a3 = [100, 110, 120, 130, 115, 110]

df = pd.DataFrame(index=x, 
                  data={'A': a1, 
                        'B': a2, 
                        'C': a3})

fig, ax = plt.subplots()
ax2 = ax.twinx()
ax3 = ax.twinx()

ax3.spines["right"].set_position(("axes", 1.1))

df['A'].plot(ax=ax, kind='bar', color='blue', width=0.2, position=2)
df['B'].plot(ax=ax2, kind='bar', color='green', width=0.2, position=1)
df['C'].plot(ax=ax3, kind='bar', color='red', width=0.2, position=0)

Answer 1

There are a couple things you can do to make this more readable. Your big issue is that you have 3 separate y-axis so that its both hard to discern which goes to which variable and you have a variable zero line (which the bars are defined from). You can help the readability first by changing the axis colors to fit your data. Then you want to set your limits for all your y-axis so that they the zero line is the same and use some multiplication factor to then adjust your scales. Be careful though because this could mislead readers dependent on the physical significance of comparing 'A' to 'B' to 'C'.

import pandas as pd
import matplotlib.pyplot as plt

x = range(6)
a1 = [-1, -4, -3, -6, 2, 8]
a2 = [ 4, 12, 8, 1, 10, 9]
a3 = [100, 110, 120, 130, 115, 110]

df = pd.DataFrame(index=x, 
                  data={'A': a1, 
                        'B': a2, 
                        'C': a3})

fig, ax = plt.subplots()
ax2 = ax.twinx()
ax3 = ax.twinx()

ax3.spines["right"].set_position(("axes", 1.1))

df['A'].plot(ax=ax, kind='bar', color='blue', width=0.2, position=2)
df['B'].plot(ax=ax2, kind='bar', color='green', width=0.2, position=1)
df['C'].plot(ax=ax3, kind='bar', color='red', width=0.2, position=0)

#Set the limits based off your negative bar graph then multiply those by some factor
ax.set_ylim(df['A'].min()*1.1,df['A'].max()*1.1) 
ax2.set_ylim(df['A'].min()*2,df['A'].max()*2)
ax3.set_ylim(df['A'].min()*20,df['A'].max()*20)

#Change color of axis to make more readable
ax.tick_params(axis='y',color='blue',labelcolor='blue')
ax2.tick_params(axis='y',color='green',labelcolor='green')
ax3.tick_params(axis='y',color='red',labelcolor='red')

#Also add a limit to the x-axis to     
ax.set_xlim(-0.5)

plt.show()