Unable to show Pandas dateindex on a matplotlib x axis

Question

I'm trying to build matplotlib charts whose x-axis is a dateIndex from a pandas dataframe. Trying to mimic some examples from matplotlib, I've been unsuccessful. The xaxis ticks and labels never appear.

I thought maybe matplotlib wasn't properly digesting the pandas index, so I converted it to ordinal with the matplotlib date2num helper function, but that gave the same result.

# https://matplotlib.org/api/dates_api.html
# https://matplotlib.org/examples/api/date_demo.html

import datetime as dt

import matplotlib.dates as mdates
import matplotlib.cbook as cbook

import matplotlib.dates as mpd


years = mdates.YearLocator()   # every year
months = mdates.MonthLocator()  # every month
yearsFmt = mdates.DateFormatter('%Y')


majorLocator = years
majorFormatter = yearsFmt #FormatStrFormatter('%d')
minorLocator = months


y1 = np.arange(100)*0.14+1
y2 = -(np.arange(100)*0.04)+12

"""neither of these indices works"""
x = pd.date_range(start='4/1/2012', periods=len(y1))
#x = map(mpd.date2num, pd.date_range(start='4/1/2012', periods=len(y1)))

fig, ax = plt.subplots()
ax.plot(x,y1)
ax.plot(x,y2)

ax.xaxis.set_major_locator(years)
ax.xaxis.set_major_formatter(yearsFmt)
ax.xaxis.set_minor_locator(months)

datemin = x[0]   
datemax = x[-1]  
ax.set_xlim(datemin, datemax)
fig.autofmt_xdate()
plt.show()

ImportanceOfBeingErnest · Accepted Answer · 2017-11-12 09:40:08Z

The problem is the following. pd.date_range(start='4/1/2012', periods=len(y1)) creates dates from the first of April 2012 to the 9th of July 2012.
Now you set the major locator to be a YearLocator. This means, that you want to have a tick for each year on the axis. However, all dates are within the same year 2012. So there is no major tick to be shown within the plot range.

The suggestion would be to use a MonthLocator instead, such that the first of each month is ticked. Also if would make sense to use a formatter, which actually shows the months, e.g. '%b %Y'. You may use a DayLocator for the minor ticks, if you want, to show the small tickmarks for each day.

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %Y'))
ax.xaxis.set_minor_locator(mdates.DayLocator())

Complete example:

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

y1 = np.arange(100)*0.14+1
y2 = -(np.arange(100)*0.04)+12

x = pd.date_range(start='4/1/2012', periods=len(y1))

fig, ax = plt.subplots()
ax.plot(x,y1)
ax.plot(x,y2)

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %Y'))
ax.xaxis.set_minor_locator(mdates.DayLocator())

fig.autofmt_xdate()
plt.show()

piRSquared · Accepted Answer · 2017-11-12 03:40:06Z

2

You could use pd.DataFrame.plot to handle most of that

df = pd.DataFrame(dict(
    y1=y1, y2=y2
), index=x)

df.plot()

answered Nov 12, 2017 at 3:40

piRSquared

296k68 gold badges509 silver badges654 bronze badges

1 Comment

user3556757 Over a year ago

thanks this is a good point. The biggest concern I have with abstracting behind pd.plot() is that it adds one more layer of control to matplotlib. I already find it often confusing wrangling matplotlib because there are various ways to treat it (as a state machine, as an object), and i tend to have those approaches commingle and cause confusion.

Collectives™ on Stack Overflow

Unable to show Pandas dateindex on a matplotlib x axis

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related