[FIXED] How to barplot time on the x-axis?

Issue

I want to create a bar plot where the x-axis is the datetime column from my data-frame.

I converted the datetime in pandas column to time only and I am getting an error when I try to use the matplotlib.pyplot.barplot function to plot the data:

import pandas as pd
import matplotlib.pyplot as plt

dataset = pd.read_csv('data.csv')
dataset['date'] = pd.to_datetime(dataset['date']).dt.time

plt.bar(dataset['date'], dataset['tickqty'])
plt.show()

I get the following error:

TypeError: unsupported operand type(s) for -: 'datetime.time' and 'float'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):

  File "<ipython-input-2-7ddd22154cf4>", line 7, in <module>
plt.bar(dataset['date'], dataset['tickqty'])

  File "/home/mahmoud/anaconda3/envs/fxenv/lib/python3.7/sitepackages/matplotlib/pyplot.py", line 2472, in bar
**({"data": data} if data is not None else {}), **kwargs)

  File "/home/mahmoud/anaconda3/envs/fxenv/lib/python3.7/site-packages/matplotlib/__init__.py", line 1431, in inner
return func(ax, *map(sanitize_sequence, args), **kwargs)

  File "/home/mahmoud/anaconda3/envs/fxenv/lib/python3.7/site-packages/matplotlib/axes/_axes.py", line 2464, in bar
f'are incompatible') from e

 TypeError: the dtypes of parameters x (object) and width (float64) are incompatible

This is the data from the csv file:

How can I plot the time column in matplotlib.pyplot.barplot?

Solution

Solution with matplotlib bar plot

pd.to_datetime(dataset['date']).dt.time converts the dates to strings (dtype object) which is causing the TypeError. For plt.bar to work, you need to use numerical data. You can use the full dates stored as pandas Timestamp objects that you get when using pd.to_datetime(dataset['date']) and matplotlib will automatically process them as matplotlib date units (number of days since 1970-01-01 UTC).

Then you need to format the x-axis tick labels so that only the times are shown. This can be done by using the matplotlib.dates DateFormatter with the appropriate format codes. Here is an example using the data you shared in the image:

import pandas as pd                # v 1.1.3
import matplotlib.pyplot as plt    # v 3.3.2
import matplotlib.dates as mdates

dataset = pd.read_csv('data.csv')
dataset['date'] = pd.to_datetime(dataset['date'])
plt.bar(dataset['date'], dataset['tickqty'], width=0.002)
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%H:%M'))
plt.show()

Note that the bar width argument is using the x variable unit (i.e. matplotlib date units) which is time in days. So 0.002 corresponds to about 2 minutes and 53 seconds.

Answered By - Patrick FitzGerald