Error bars not rendered correctly

[drorata]

Code Sample, a copy-pastable example if possible

From the /doc/source/visualization.rst:

import pandas as pd
import matplotlib.pyplot as plt

# Generate the data
ix3 = pd.MultiIndex.from_arrays([['a', 'a', 'a', 'a', 'b', 'b', 'b', 'b'], ['foo', 'foo', 'bar', 'bar', 'foo', 'foo', 'bar', 'bar']], names=['letter', 'word'])
df3 = pd.DataFrame({'data1': [3, 2, 4, 3, 2, 4, 3, 2], 'data2': [6, 5, 7, 5, 4, 5, 6, 5]}, index=ix3)

# Group by index labels and take the means and standard deviations for each group
gp3 = df3.groupby(level=('letter', 'word'))
means = gp3.mean()
errors = gp3.std()
means
errors

# Plot
fig, ax = plt.subplots()
means.plot.bar(yerr=errors, ax=ax)

Problem description

The resulting image doesn't contain the horizontal ticks of the error bars.

Expected Output

To achieve this the last line should be replaced with

means.plot.bar(yerr=errors, ax=ax, capsize=4)

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.2.final.0
python-bits: 64
OS: Darwin
OS-release: 16.7.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: en_US.UTF-8
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.20.3
pytest: 3.2.1
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.26
numpy: 1.13.1
scipy: 0.19.1
xarray: None
IPython: 6.1.0
sphinx: 1.6.3
patsy: 0.4.1
dateutil: 2.6.1
pytz: 2017.2
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.0.2
openpyxl: 2.4.8
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: 0.9.8
lxml: 3.8.0
bs4: 4.6.0
html5lib: 0.999
sqlalchemy: 1.1.13
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
pandas_gbq: None
pandas_datareader: None

[justinchan23]

I will take a stab at this!

[TomAugspurger]

I don't recall perfectly, but I think in in #17468 I expressed concern with deviating from the matplotlib default style.

[dat-boris]

Thanks @TomAugspurger !

Ah in #17468 - we saw an issue of unrelated code getting pulled in and thought that's why the PR was aborted. Thanks for the extra context.

So to summarize from #17468 there are a few suggestions:

From @TomAugspurger 👍

, Either way, I think aesthetically the version without caps looks better. What do you think?

From @drorata

As for the aesthetic aspect --- I think the caps are crucial for faster and better perception of the error bars.

I'm really dont want to get back into the "passionate" debate about error cap as have others previously done in matplotlib/matplotlib#5047 😅

I am happy if this goes either way - I think there's pros and cons on both sides as already demonstrated as in the above discussion.

Although I do have a slight preference in demonstrating that capability in the documentation.

My thought is that it helps to show an example of capsize keyword argument and allow a developer to choose the right way to go. I don't think the documentation is representative of suggested styling.

Let's decide and either:"

1. close this issue
2. merge the PR?

Thank you!
Boris

[TomAugspurger]

I think an example in visualization.rst showing how capsize works would be valuable. I may have been misremembering the scope of the old PR. Taking a look at #23674

…

On Tue, Nov 13, 2018 at 2:58 PM Boris Lau ***@***.***> wrote: Thanks @TomAugspurger <https://github.com/TomAugspurger> ! Ah in #17468 <#17468> - we saw an issue of unrelated code getting pulled in and thought that's why the PR was aborted. Thanks for the extra context. So to summarize from #17468 <#17468> there are a few suggestions: From @TomAugspurger <https://github.com/TomAugspurger> 👍 , Either way, I think aesthetically the version without caps looks better. What do you think? From @drorata <https://github.com/drorata> As for the aesthetic aspect --- I think the caps are crucial for faster and better perception of the error bars. I'm really dont want to get back into the "passionate" debate about error cap as have others previously done in matplotlib/matplotlib#5047 <matplotlib/matplotlib#5047> 😅 I am happy if this goes either way - I think there's pros and cons on both sides as already demonstrated as in the above discussion. Although I do have a slight preference in demonstrating that capability in the documentation. My thought is that it helps to show an example of capsize keyword argument and allow a developer to choose the right way to go. I don't think the documentation is representative of suggested styling. Let's decide and either:" 1. close this issue 2. merge the PR? Thank you! Boris — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#17467 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABQHImixzA5EA4Umw7bWNaduiKlbR4CLks5uuzKIgaJpZM4PQRDC> .