如何将字符串从 csv 转换为小时：分钟？

Question

text

If I copied well the link, you two picture. One shows the csv file I have and the other the graph I want to create.

It would be easy so there is a but,

I want to represent the AVG number (which are seconds actually) as hour(s):minute(s) on y axis.

I think, it cannot be solved because I spent 3 days wit this problem.

Everything, in terms of my googling skills and etc...

But to be more precise, aside of lot of conversations with dateime, timedelta, timestamp nothing worked.

Either the data could no be shown on y axis because it did not represent number like variable to plot or I've got not proper representation of the data.

I was trying to create a different approach, like convert the seconds calculating with divmod than put them on the top of the bars with annonate. There are problem with this, still...

Later I was thinking I have found the holy grail:

text

I could not use this, I do not understand how should I create an acceptable datatype for this..

Thanks for anyone who can help!!

Answer 1

the example you show in the first image could be reproduced by using plot.bar from pandas

>>> import pandas as pd
>>> df = pd.DataFrame()
>>> df["activity"] = ['run', 'swim', 'drive']
>>> df["avg"] = [86400,43200,21600]
>>> df
  activity    avg
0      run  86400
1     swim  43200
2    drive  21600
>>> df.plot.bar(x="activity")
<AxesSubplot: xlabel='activity'>
>>> plt.show()

To represent time transcurred for a certain number of seconds you can use fromtimestamp and formatting strftime but it might not be compatible with matplotlib, then using Timple is something related, but graph could not be properly plotted or maybe it is needed to perform something like exploratory data analysis or apply a certain statistical procedure.

>>> import datetime
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> import timple
>>>
>>> tmpl = timple.Timple()
>>> tmpl.enable()
>>> timedeltas = np.array([datetime.timedelta(seconds=(s)) for s in df["avg"]])
>>> timedeltas
array([datetime.timedelta(days=1), datetime.timedelta(seconds=43200),
       datetime.timedelta(seconds=21600)], dtype=object)
>>> plt.plot(timedeltas, df["activity"])
[<matplotlib.lines.Line2D object at 0x0000026FAC3F5B40>]
>>> plt.show()