How to create a scatter plot with timestamps

Question

How to produce a scatter plot using timestamps?

Below is an example but I get an error ValueError: First argument must be a sequence

import pandas as pd
import matplotlib.pyplot as plt

d = ({
    'A' : ['08:00:00','08:10:00'],
    'B' : ['1','2'],           
    })

df = pd.DataFrame(data=d)

fig = plt.figure()

x = df['A']
y = df['B']

plt.scatter(x,y)

If I convert the timestamps to total seconds it works:

df['A'] = pd.to_timedelta(df['A'], errors="coerce").dt.total_seconds()

But I'd like 24hr time to be on the x-axis, rather than total seconds.

Tim Skov Jacobsen · Accepted Answer · 2021-11-09T11:21:02.843

2

Use x-ticks:

d = ({
'A' : ['08:00:00','08:10:00'],
'B' : [1,2],
})

df = pd.DataFrame(data=d)

fig = plt.figure()

x = df['A']
y = df['B']

x_numbers = list(pd.to_timedelta(df['A'], errors="coerce").dt.total_seconds ())
plt.scatter(x_numbers, y)
plt.xticks(x_numbers, x)
plt.show()

This will plot the timestamps as the tick marks on the x-axis.

Note that I converted the 'B' column in your dataframe to numbers instead of strings. If they really need to be strings or are imported as strings from a file, just convert them to numbers, e.g. via float() or int().

edited Nov 09 '21 at 11:21

answered Jun 22 '18 at 06:56

Tim Skov Jacobsen

3,583
4
26
23

Thanks @Tim Skov Jacobsen. It works without converting to lists – Jun 23 '18 at 01:23
@Punter345. You are right. When plotting, it is not necessary. I have removed it from the answer. – Tim Skov Jacobsen Jun 25 '18 at 12:29

How to create a scatter plot with timestamps

1 Answers1