CODEWITHSUNDEEP
pythonpandasplotmulti-index
user2471066
user2471066

Reputation: 41

Plotting multi indexed data with pandas

I have a Python dictionary in the form of:

dict = {0: {a: 1, b: 2}, 1: {a: 1, b: 2}}

I created a Pandas DataFrame from it, by using from_dict(). This gave me a multiindex dataframe`, as expected.

Now I want to plot this dataframe with the outer index being the x-axis and the inner index being separate lines in my graph, the data points on the y-axis being the values.

I tried stacking them, but it only gave me a single line.

My current code:

data = pd.DataFrame.from_dict(my_data)
data = data.T.stack()
data.plot()

Anyone has an idea how this could be done?

Upvotes: 1

Views: 1183

Answers (1)

unutbu
unutbu

Reputation: 879421

Use 

data.groupby(level=1).plot(stacked=True)

import pandas as pd
import matplotlib.pyplot as plt

my_data = {0: {'a': 1, 'b': 2}, 1: {'a': 1, 'b': 2}}
data = pd.DataFrame.from_dict(my_data)
data = data.T.stack()
# 0  a    1
#    b    2
# 1  a    1
#    b    2
# dtype: int64

data.groupby(level=1).plot(stacked=True, legend=True)
plt.show()

Or, for more control in configuring the plot it may be easier to use matplotlib directly:

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

data = pd.concat({key:pd.Series(np.random.randint(10, size=(10,))) for key in 'AB'}).swaplevel()

fix, ax = plt.subplots()
for key, grp in data.groupby(level=1):
    ax.plot(grp.index.get_level_values(0), grp.values, label=key)
plt.legend()
plt.plot()
plt.show()

enter image description here

Upvotes: 4

Related Questions