Python matplotlib/seaborn plot bar chart subcategories

Question

I have a dictionary of dictionaries called data. I want to plot a bar chart so that each of A, B, C, and D have a value for the pos and neg.

I don't want the pos and neg bars to be stacked on top of each other. I want them side by side.

Additionally, I want the categories to be sorted by descending order of their total frequency (pos + neg)

data = {'A': {'pos': 289794, 'neg': 515063},
        'B': {'pos': 174790, 'neg': 292551},
        'C': {'pos': 375574, 'neg': 586616},
        'D': {'pos': 14932, 'neg': 8661}}

What's your question? Do you face any problems when doing this kind of chart? — ImportanceOfBeingErnest, Sep 01 '19 at 14:20
@ImportanceOfBeingErnest The main issue for me is that I want each category for my bar chart to have two subcategories (`pos` and `neg`) — piccolo, Sep 01 '19 at 14:24
That would be easiest solved by `pd.DataFrame(data).T.plot.bar()` I suppose? — ImportanceOfBeingErnest, Sep 01 '19 at 14:30
@ImportanceOfBeingErnest this works. Can you add it as an answer so that I can mark it as the correct one. Also shouldn't `matplotlib` have an inbuilt function to do this since it seems like situations like this would be common — piccolo, Sep 01 '19 at 14:35
[This is the second example in the matplotlib gallery](https://matplotlib.org/3.1.1/gallery/lines_bars_and_markers/barchart.html) — ImportanceOfBeingErnest, Sep 01 '19 at 14:43

ComplicatedPhenomenon · Accepted Answer · 2019-09-01T14:56:11.257

2

%matplotlib inline 
import matplotlib.pyplot as plt
import pandas as pd

data = {'A': {'pos': 289794, 'neg': 515063},
        'B': {'pos': 174790, 'neg': 292551},
        'C': {'pos': 375574, 'neg': 586616},
        'D': {'pos': 14932, 'neg': 8661}}
df = pd.DataFrame(data)
df = df.T
df ['sum'] = df.sum(axis=1)
df.sort_values('sum', ascending=False)[['neg','pos']].plot.bar()

edited Sep 01 '19 at 14:56

answered Sep 01 '19 at 14:49

ComplicatedPhenomenon

4,055
2
18
45

Python matplotlib/seaborn plot bar chart subcategories

1 Answers1