CODEWITHSUNDEEP
pythonpandastime-series
TMrtSmith
TMrtSmith

Reputation: 471

Combining time series into single plot with different time instances

I have a large dataframe like the one below, from which I want to plot a time series showing a line for the temperature from each sender over time.

                utctime      sender  temp_3
500 2014-10-24 11:21:08           e    24.7
501 2014-10-24 11:21:09           d    22.8
502 2014-10-24 11:21:09           a    23.2
503 2014-10-24 11:21:09           c    24.3
504 2014-10-24 11:21:10           b    23.9
505 2014-10-24 11:21:10           e    24.7
506 2014-10-24 11:21:11           d    22.9
507 2014-10-24 11:21:11           a    23.1
508 2014-10-24 11:21:11           c    24.2
509 2014-10-24 11:21:12           b    23.9
510 2014-10-24 11:21:12           e    24.7
511 2014-10-24 11:21:13           d    22.9
512 2014-10-24 11:21:13           a    23.1
513 2014-10-24 11:21:13           c    24.2
514 2014-10-24 11:21:14           b    23.9

I tried using filtering to pull out a Series for each sender temps, and then recombining these into a new dataframe, but the times are all different. Is there another way to do this? I'm new to this, so apologise if this is a repeat!

Upvotes: 1

Views: 129

Answers (1)

fernandezcuesta
fernandezcuesta

Reputation: 2448

Use pandas.DataFrame.groupby() method to group by sender, then plot. As an example:

plotaxis = plt.figure().gca()
for key, grp in dataframe.groupby(['sender']):
    my_ts = [ts.to_julian_date() - 1721424.5
             for ts in grp['utctime'].dropna()]
    plt.plot(my_ts,
             grp['temp_3'].dropna(),
             label='%s@%s' % (Temperature, key))
# Style the resulting plot
plotaxis.xaxis.set_major_formatter(
    matplotlib.dates.DateFormatter('%d/%m/%y\n%H:%M')
)

Upvotes: 1

Related Questions