Improving simple bar chart using matplotlib

Question

I wanted to create a bar chart from a employee survey using matplotlib. After following the directions on the following question

How to create a bar chart/histogram with bar per discrete value?

I came up with the code

satisfaction = survey['satisfaction'].value_counts().sort_index()
ax = satisfaction.plot(kind='bar')
fig = ax.get_figure()
fig.autofmt_xdate()
plt.show()

Which printed the following graph

bar chart of Satisfaction

So, from here.

1-How do I centralize the chart? I mean, I have a big space between the left border and the 1st column and no border at all between the last column and the right border. I want to have both spaces of the same size.

2-How do I change the columns labels? the "1" on the 1st column really means "Very Satisfied", the "2" mean "Satisfied" and so on. I want to put the word meanings instead of the numeric values.

3-How to I put the chart title and x and y axes labels.

side note: no need to call fig.autofmt_xdate(). you don't have date — Paul H
– Paul H, Commented Sep 21, 2014 at 1:08

Paul H · Accepted Answer · 2014-09-21 01:06:48Z

1

ax.set_xlim(left=..., right=...)
Probably simplest to have those values in our dataframe and plot against those
ax.set_xlabel(...) and ax.set_ylabel(...)

Note that 1 and 3 are demonstrated in most basic matplotlib tutorials.

Here's a good one: http://jakevdp.github.io/mpl_tutorial/tutorial_pages/tut2.html

For number 2, it's as simple as:

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

datamap = {
    1: 'real bad',
    2: 'bad',
    3: 'meh',
    4: 'good',
    5: 'way good'
}


df = pandas.DataFrame(data=np.random.choice(range(1,6), size=37), columns=['score'])
df['rating'] = df.score.map(datamap.get)

fig, ax = plt.subplots()
df.rating.value_counts().plot(kind='bar', ax=ax)
## alternatively:
# df.groupby(by='rating').count().plot(kind='bar', ax=ax)
fig.tight_layout()

enter image description here

edited Sep 21, 2014 at 1:06

answered Sep 19, 2014 at 16:48

Paul H

68.7k23 gold badges165 silver badges139 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

user3724295 Over a year ago

For question 2. in "R" I can get the same chart by by simple adding the argument 'names=c("my_text1","my_text2),..' to the bar chart code. It would be to cumbersome to having change all the data every time I just need a aesthetic change on the labels.

Paul H Over a year ago

@user3724295 it's real not cumbersome at all. just different. see my edits with a short, self-contained, example.

user3724295 Over a year ago

Your method works, Thanks. However, I found that ax.set_xticklabels (('my_fancy_names')) can also do the trick.

Paul H Over a year ago

@user3724295 the problem with that is that it's now your responsibility to keep the tick labels synchronized with the x-axis values and the dataframe. I prefer to make the computer do all of that work.

Collectives™ on Stack Overflow

Improving simple bar chart using matplotlib

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related