Plot chart with breakdown in pandas
I have a DataFarme df in the following form. I want to plot a chart with 4 pair bars. This means that for each of the four days, there are two columns representing 0 and 1. For both of the two bars, I want to have a stacked bar. Each bar has 3 colors representing <30, 30-70 and> 70. I tried to use "stacked = True", but instead of paired bars, one panel with 6 colors was built. Can you help someone? Many thanks.
Score <30 30-70 >70
Gender 0 1 0 1 0 1
2017-07-09 23 10 25 13 12 21
2017-07-10 13 14 12 14 15 10
2017-07-11 24 25 10 15 20 15
2017-07-12 23 17 20 17 18 17
+3
source to share
1 answer
You can use a parameter bottom
. Here's the way:
>> import matplotlib.pyplot as plt
>> import numpy as np
>> import pandas as pd
>>
>> columns = pd.MultiIndex.from_tuples([(r, b) for r in ['<30', '30-70', '>70']
>> for b in [0, 1]])
>> index = ['2017-07-%s' % d for d in ('09', '10', '11', '12')]
>> df = pd.DataFrame([[23,10,25,13,12,21], [13,14,12,14,15,10],
>> [24,25,10,15,20,15], [23,17,20,17,18,17]],
>> columns=columns, index=index)
>>
>> width = 0.25
>> x = np.arange(df.shape[0])
>> xs = [x - width / 2 - 0.01, x + width / 2 + 0.01]
>> for b in [0, 1]:
>> plt.bar(xs[b], df[('<30', b)], width, color='r')
>> plt.bar(xs[b], df[('30-70', b)], width, bottom=df[('<30', b)], color='g')
>> plt.bar(xs[b], df[('>70', b)], width, bottom=df[('<30', b)] + df[('30-70', b)], color='b')
>> plt.xticks(x, df.index)
>> plt.show()
+2
source to share