Matplotlib Line vs. Bar plot DateTime axis formatting

Question

I have a DataFrame with a DateTime index:

import pandas as pd
from random import randrange
dates = pd.date_range(start="2020-02-01",end='2020-04-18',freq='1d')
df = pd.DataFrame(index=dates,data=[randrange(10000) for i in range(78)]

Now when I plot the data as a line plot, matplotlib produces a nicely formatted x axis:

df.plot(figsize=(12,4))

However, if instead I do a bar plot, I now get something quite horrible:

df.plot(kind='bar',figsize=(12,4)),

This is quite disconcerting, as it is the same DataFrame. What I want is to have the bar plot, but with the nicely formatted DateTime axis.

Is there an easy way to achieve that? I tried using the bar command directly, but that produces yet another x axis format, not as nice as the one generated by plotting with a pandas method.

This should help you :https://matplotlib.org/gallery/api/date_index_formatter.html — Rajat Mishra, Apr 19 '20 at 04:57
@Rajat: Thanks. Yes, I understand that I can manually pick a DateTime format. There are two problems with this: First, it requires a lot of work. Second, how would I generate the same nice format that I get with the line plot? — germ, Apr 19 '20 at 17:21
Is there a way to take the axis generated by the line plot and re-use it for the bar plot? I tried doing that (by copying locations and formats from xticks), but I was not successful. I think the reason is that the ticks structure for the bar plot is different than the line plot. — germ, Apr 19 '20 at 17:26
May be this could help https://stackoverflow.com/questions/30133280/pandas-bar-plot-changes-date-format — Rajat Mishra, Apr 19 '20 at 17:55

score 1 · Answer 1 · answered Apr 19 '20 at 05:08

You can do something like this :

In [23]: fig, ax = plt.subplots(figsize=(12, 12))
Attribute Qt::AA_EnableHighDpiScaling must be set before QCoreApplication is created.

In [27]: ax.bar(df.index.values,
    ...:        df[0],
    ...:        color='purple')
Out[27]: <BarContainer object of 78 artists>

In [29]: import matplotlib.dates as mdates
    ...: from matplotlib.dates import DateFormatter

In [30]: date_form = DateFormatter("%m-%d")
    ...: ax.xaxis.set_major_formatter(date_form)
    ...:
    ...: # Ensure a major tick for each week using (interval=1)
    ...: ax.xaxis.set_major_locator(mdates.WeekdayLocator(interval=1))
    ...: plt.show()

Yes, that is about how far I got as well. It's still not as nice as the one from the line plot... — germ, Apr 19 '20 at 17:22

Matplotlib Line vs. Bar plot DateTime axis formatting

1 Answers1

Linked