Converting datetime.time to int for use in regression model / initial scatter plot

Question

I am looking to change a datetime.time column in a pandas dataframe to an int value that isnt a unix like time i.e. from 1970 epoch.

Example of df.head(3):

trans_date_trans_time	amt	category	city_pop	is_fraud	Time of Day
2019-01-01 00:00:18	4.97	misc_net	3495	0	00:00:18
2019-01-01 00:00:44	107.23	grocery_pos	149	0	00:00:44
2019-01-01 00:00:51	220.11	entertainment	4154	0	00:00:51

So i want the 'Time of Day' column to read like an integer that can run along an axis of a simple scatter plot against 'amt'.

When I try:

y = int(df_full_slim['Time of Day'])

plt.scatter(x, y)
plt.show() 

or simply:

y = df_full_slim['Time of Day']

plt.scatter(x, y)
plt.show()

it doesnt work as it cant plot a datetime.time type on a plt.

How do I get the time in a format that will run along an axis of a plot?

Thanks in advance..

Answer 1

You can plot without conversion by calling the .plot() method of the dataframe:

df_full_slim.plot(x='Time of Day', y='amt')