如何从熊猫系列或数据框计算最大15分钟的总和

Question

Pandas newbie here. I have a dataset which contains traffic counts with time stamps. I want to know which 15-min interval has the most cumulative sum of counts, and the value of this sum.

Data might look something like this:

import random 
ts = pd.Series(range(1000),index=random.sample(pd.date_range('2015-02-01 06:00:00',periods=3000,freq='1min'),1000)).sort_index()

2015-02-01 06:06:00    314
2015-02-01 06:08:00    154
2015-02-01 06:09:00    914
2015-02-01 06:13:00     84
2015-02-01 06:18:00    880
2015-02-01 06:22:00    912
2015-02-01 06:28:00    410
2015-02-01 06:32:00    391
2015-02-01 06:34:00    270
2015-02-01 06:35:00    984
2015-02-01 06:36:00    271
2015-02-01 06:37:00    722
2015-02-01 06:38:00    748
2015-02-01 06:40:00    313
2015-02-01 06:42:00    277
2015-02-01 06:43:00    604
2015-02-01 06:49:00    888
2015-02-01 06:50:00    943
2015-02-01 06:51:00    124
2015-02-01 06:52:00    806

Is there a way to do this in Pandas?

Answer 1

a simple solution without using pandas native functions

from datetime import timedelta
start = ts.index[0]
end = ts.index[len(ts)-1]
dur = timedelta(minutes=15)
max_val = 0
while start < end:
    cum_sum = ts[start : start+dur].sum()
    if cum_sum > max_val:
        max_val = cum_sum
        max_seg = (start, start+dur)
    start = star+dur 
print max_val
print max_seg

Answer 2

This is what I came up with:

def find_peak_15_minutes(data_frame, column):

    max_sum = 0
    start_of_max15 = 0
    for start in data_frame[column].values:
        series_sum = data_frame[column][data_frame[column].between(start, start + 15)].count()
        if series_sum > max_sum:
            max_sum = series_sum
            start_of_max15 = start
    return (start_of_max15, max_sum)