How can I convert pandas date time xticks to readable format?

Question

I am plotting a time series with a date time index. The plot needs to be a particular size for the journal format. Consequently, the sticks are not readable since they span many years.

Here is a data sample

2013-02-10  0.7714492098202259
2013-02-11  0.7709101833765016
2013-02-12  0.7704911332770049
2013-02-13  0.7694975914173087
2013-02-14  0.7692108921323576

The data is a series with a datetime index and spans from 2013 to 2016. I use

data.plot(ax = ax)

to plot the data.

How can I format my xticks to read like '13 instead of 2013?

FYI- you can also consider using `plt.tight_layout()`. This should adjust and fit the x labels. — Andrew L, Jun 19 '17 at 19:03

score 1 · Accepted Answer · answered Jun 20 '17 at 08:59

It seems there is some incompatibility between pandas and matplotlib formatters/locators when it comes to dates. See e.g. those questions:

I'm not entirely sure why it still works in some cases to use matplotlib formatters and not in others. However because of those issues, the bullet-proof solution is to use matplotlib to plot the graph instead of the pandas plotting function. This allows to use locators and formatters just as seen in the matplotlib example.

Here the solution to the question would look as follows:

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import numpy as np

dates = pd.date_range("2013-01-01", "2017-06-20" )
y = np.cumsum(np.random.normal(size=len(dates)))

s = pd.Series(y, index=dates)

fig, ax = plt.subplots()
ax.plot(s.index, s.values)

ax.xaxis.set_major_locator(mdates.YearLocator())
ax.xaxis.set_minor_locator(mdates.MonthLocator())
yearFmt = mdates.DateFormatter("'%y")
ax.xaxis.set_major_formatter(yearFmt)

plt.show()

Ted Petrou · Answer 2 · 2017-06-20T18:57:33.113

0

According to this example, you can do the following

import matplotlib.dates as mdates

yearsFmt = mdates.DateFormatter("'%y")
years = mdates.YearLocator() 
ax = df.plot()
ax.xaxis.set_major_locator(years)
ax.xaxis.set_major_formatter(yearsFmt)

Full work below

Add word value so pd.read_clipboard puts dates into index

value
2013-02-10  0.7714492098202259
2014-02-11  0.7709101833765016
2015-02-12  0.7704911332770049
2016-02-13  0.7694975914173087
2017-02-14  0.7692108921323576

Then read in data and convert index

df = pd.read_clipboard(sep='\s+')
df.index = pd.to_datetime(df.index)

edited Jun 20 '17 at 18:57

answered Jun 19 '17 at 19:09

Ted Petrou

59,042
19
131
136

After trying this, this is writing the year somewhere in the forties. – Demetri Pananos Jun 19 '17 at 19:12
I changed your data to different years and this is the image I am getting. – Ted Petrou Jun 19 '17 at 19:16
The output from the code above is [something like this](https://i.stack.imgur.com/oj90e.png) for me, with completely wrong years. @Ted can you provide a full working code that produces the plot you show? – ImportanceOfBeingErnest Jun 20 '17 at 08:53
@ImportanceOfBeingErnest I just put the first column in the index and coverted it to a pandas timestamp. Could this be an issue with the date directive? I'm on a mac. – Ted Petrou Jun 20 '17 at 18:47
I don't know. To find out, it would be good to show your complete code. This would possibly provide an acceptable solution to this problem. – ImportanceOfBeingErnest Jun 20 '17 at 18:50

How can I convert pandas date time xticks to readable format?

2 Answers2

Linked