4.2. Rules of Object Oriented Matplotlib

import pandas as pd
import matplotlib.pyplot as plt

4.2.1. Example 1 - Without defining Axis

• If you want to be quick and do not want to define axis manually and just want to do ax.plot(df), you will need to structure your df in specific way

  • By default, index is taken as x axis

  • So, you will need to change the index of your df to whatever column you want as x axis

  • There should be no other categorical variable

  • All the numerical columns will be plotted as separate lines

data = {'Month':['Jan','Feb','March','April'],
        'Sales': [99, 98, 95, 90],
        'Profit': [10,20,30,40]
       }
df=pd.DataFrame(data,columns=['Sales','Profit'],index=data['Month'])
df

# See how I changed the structure of dataframe in pd.dataframe

	Sales	Profit
Jan	99	10
Feb	98	20
March	95	30
April	90	40

fig,ax=plt.subplots()
ax.plot(df)
ax.set_title('Product Sales vs Profits')

Text(0.5, 1.0, 'Product Sales vs Profits')

4.2.2. Example 2 - Manually defining Axis

data = {'Subject': ['A', 'B', 'C', 'D'],
        'SahilMarks': [99, 98, 95, 90],
        'SoniaMarks': [10,20,30,40]
       }
df=pd.DataFrame(data)
df

	Subject	SahilMarks	SoniaMarks
0	A	99	10
1	B	98	20
2	C	95	30
3	D	90	40

fig,ax=plt.subplots()
ax.plot(df.index,df['SahilMarks']) # Manually giving x axis
ax.plot(df.index,df['SoniaMarks'],ls='--') # Manually giving y axis

# Note, if you don't provide axis, directly use ax.plot(df) ->
    # Index becomes x axis
    # remaining number column becomes y axis
    # Fails if more than one number column so its better to define manually

[<matplotlib.lines.Line2D at 0x7f8a8c19b9d0>]

4.2.3. Summary

• If you want quick analysis, pass entire df to ax.plot()

• If you want customizations of each line, declare them separately and plot each of them separately using ax.plot() multiple times

• Pro Tip: Lets say you want different line for each country then you will need different column for different countries (since lines are columns in matplotlib) so pivoting is required for that and keep index as your only dimension (Keep only One dimension)