How to do draw a seaborn or a matlab plot

Question

I have a dataframe which looks like below

            A   B   C
Above 90%   30  50  60
80% - 90%   39  72  79
65% - 80%   80  150 130
Below 65%   80  70  90

And I would like to get a bar chart like below

And I have another dataset which has one more sub category for A, B and C

Company Category    Above 90%   80% - 90%   65% - 80%   Below 65%
A       Fast Movers 1           11          21          9
A       Med Movers  2           13          42          40
A       Slow Movers 34          26          44          50
B       Fast Movers 1           11          21          9
B       Med Movers  16          48          92          45
B       Slow Movers 33          13          37          45
C       Fast Movers 0           0           2           30
C       Medi Movers 11          37          74          50
C       Slow Movers 41          42          65          30

For the above table I want a stacked bar chart with colors only for categories and I want Company and accuracy to be labels not colors.

I was not even able to achieve it using other softwares.

I request you all to kindly help me to achieve it, I will appreciate your time and help.

Thank you in advance

Please share the code of what you tried so far and tell us what error you get. — Patrick Artner, May 29 '18 at 06:23

score 3 · Accepted Answer · answered May 29 '18 at 06:35

3

For the first question, you can use this:

import matplotlib.pyplot as plt
ax = df.plot(kind='bar', title ="Accuracy distribution", figsize=(15, 10), legend=True, fontsize=12)
for p in ax.patches:
    ax.annotate(str(p.get_height()), (p.get_x() * 1.005, p.get_height() * 1.005))
plt.show()

answered May 29 '18 at 06:35

Joe

12,057
5
39
55

Hi Joe, thanks for answering. What if I want to display the xlabels horizontally , not vertically – Ahamed Moosa May 29 '18 at 07:11
@AhamedMoosa You are welcome. Add this line of code `plt.xticks(rotation='horizontal')` – Joe May 29 '18 at 08:35
Hi joe, is there anyway that you can use to display the lables in int rather than float – Ahamed Moosa May 29 '18 at 11:03
@AhamedMoosa add `int` like here: `ax.annotate(str(int(p.get_height())), (p.get_x() * 1.005, p.get_height() * 1.005))` – Joe May 29 '18 at 11:06

score 2 · Answer 2 · answered May 29 '18 at 06:36

2

You can use:

df.set_index(['Company','Category']).stack().unstack(1).plot.bar(stacked=True)

Or for each company separate graph:

for i, x in df.set_index(['Company','Category']).stack().unstack(1).reset_index(level=0).groupby('Company'):
    x.plot.bar(stacked=True)

answered May 29 '18 at 06:36

jezrael

822,522
95
1,334
1,252

Thanks Jezrael, I used df.set_index(['Accuracy']).plot.bar(stacked=False, figsize=(10, 5)) for my first chart. Is there anyway I can include the data labels and resize it – Ahamed Moosa May 29 '18 at 07:04
@AhamedMoosa - Not easy, i think need [`this`](https://stackoverflow.com/a/42498711) solution – jezrael May 29 '18 at 07:40
1

Noted mate, Thanks for all your help. I really appreciate it – Ahamed Moosa May 29 '18 at 10:51

How to do draw a seaborn or a matlab plot

2 Answers2