converting the duration column into short, medium, long values

user_19555780 0
Viewed 35

how to Convert the duration column into short, medium, long values. Come up with the boundaries by splitting the duration range in 3 equal size ranges

df['duration'].head()

Output:

0      01:13
1      01:09
2      01:40
3      00:48
4      00:49
       ...  
98     00:52
99     01:10
100    01:05
101    01:09
102    00:58
Name: duration, Length: 103, dtype: object
binning pandas dataframe python
1 Answers

You can use pandas.cut on the Series of Timedelta after conversion with pandas.to_timedelta:

df['bin'] = pd.cut(pd.to_timedelta(df['duration']+':00'),
                   bins=3, labels=['low', 'medium', 'high'])

NB. assuming a HH:MM format here, for MM:SS, use '00:'+df['duration']

output:

    duration     bin
0      01:13  medium
1      01:09  medium
2      01:40    high
3      00:48     low
4      00:49     low
...      ...     ...
98     00:52     low
99     01:10  medium
100    01:05     low
101    01:09  medium
102    00:58     low

If you also want the boundaries:

df['bin'], boundaries = pd.cut(pd.to_timedelta(df['duration']+':00'),
                               bins=3, labels=['low', 'medium', 'high'],
                               retbins=True
                              )

print(boundaries)

output:

TimedeltaIndex(['0 days 00:47:56.880000',        '0 days 01:05:20',
                       '0 days 01:22:40',        '0 days 01:40:00'],
               dtype='timedelta64[ns]', freq=None)
user_163434640
Related