Search code examples
pythondatetimematplotlibformattingxticks

Problem with plotting data using datetime format for xticks

I have a time series that I obtain from a numerical simulation with an integration time step of 1min. This time series spans several months and I want to use dates in the x-axis. Once the integration is over I add an epoch time corresponding with the first day to the time numpy array created as

tt = np.arange(0.0, number_of_days, minute)

I only pick up data points every 10 mins (tt[::10]). I convert this array to the date format '%Y-%m-%d %H:%M:%S' via

dtime = np.empty(len(tt[::10]), dtype="S19")
for ind, t in enumerate(tt[::10]):
   daytime = time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(t))
   dtime[ind] = datetime.strptime(daytime, '%Y-%m-%d %H:%M:%S')

print(dtime) results in:

[b'2020-01-22 00:00:00' b'2020-01-22 00:10:00' b'2020-01-22 00:20:00' ...
 b'2020-06-29 23:30:00' b'2020-06-29 23:40:00' b'2020-06-29 23:50:00']

I format the xticks as:

plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%Y-%m-%d'))
plt.gca().xaxis.set_major_locator(mdates.WeekdayLocator())
plt.gca().xaxis.set_minor_locator(mdates.DayLocator())

I plot this data with:

plt.plot_date(dtime, Y)
plt.gcf().autofmt_xdate()

I get the same result if I only use plt.plot. I get no xticks with the dates unfortunately. It seems dates are not properly formatted in datetime format so that matplotlib understands it. I suspect the problem has to do with dtime, which I think should not be of string type. The result is

enter image description here

I would like to plot this time series with this graph, which has one data point per day.

enter image description here

I have a second related question here: how to set the major ticks at biweekly intervals?

When I plot both datasets together I get this:

enter image description here

Update: Based on the solution proposed by Andrea, I can now plot both datasets in the same graph as below

enter image description here

Solution

  • Why do not you use pd.date_range to generate the dtime array?

    dtime = pd.date_range(start = '2020-01-22', end = '2020-06-29', freq = '10min')

    In this way, the dtime is:

    DatetimeIndex(['2020-01-22 00:00:00', '2020-01-22 00:10:00',
               '2020-01-22 00:20:00', '2020-01-22 00:30:00',
               '2020-01-22 00:40:00', '2020-01-22 00:50:00',
               '2020-01-22 01:00:00', '2020-01-22 01:10:00',
               '2020-01-22 01:20:00', '2020-01-22 01:30:00',
               ...
               '2020-06-28 22:30:00', '2020-06-28 22:40:00',
               '2020-06-28 22:50:00', '2020-06-28 23:00:00',
               '2020-06-28 23:10:00', '2020-06-28 23:20:00',
               '2020-06-28 23:30:00', '2020-06-28 23:40:00',
               '2020-06-28 23:50:00', '2020-06-29 00:00:00'],
              dtype='datetime64[ns]', length=22897, freq='10T')

    With this code:

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

dtime = pd.date_range(start = '2020-01-22', end = '2020-06-29', freq = '10min')
X = np.linspace(-10, 10, len(dtime))
Y = np.exp(X) / (np.exp(X) + 1)

plt.plot(dtime, Y)

plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%Y-%m-%d'))
plt.gca().xaxis.set_major_locator(mdates.DayLocator(interval = 7))
plt.gca().xaxis.set_minor_locator(mdates.DayLocator(interval = 1))
plt.setp(plt.gca().xaxis.get_majorticklabels(), rotation = 45 )
plt.gca().set_xlim(dtime[0], dtime[-1])

plt.show()

    I get this plot:

    enter image description here

    Since I do not have access to your Y data, I generated a sigmoid function in order to make the plot; replace it with your data.