I have a set of data that has a bunch of start and end times, Using pandas i'm looking for the best way to get the number of overlapping start and end times, meaning the maximum start and end times that happen.
The table is of sessions with start and end time, the idea is to create a line graph showing the maximum concurrency of sessions over time.
import numpy as np
import pandas as pd
ranges = np.asarray([(np.datetime64('2013-12-24T00:00:00Z'), np.datetime64('2013-12-24T00:02:00Z')),
(np.datetime64('2013-12-24T00:02:00Z'), np.datetime64('2013-12-24T00:04:00Z')),
(np.datetime64('2013-12-24T00:03:00Z'), np.datetime64('2013-12-24T00:05:00Z'))])
us = pd.Series(index=np.unique(ranges), data=0)
for start, end in ranges:
us[start:end] += 1
print us
Output:
2013-12-24 00:00:00 1
2013-12-24 00:02:00 2
2013-12-24 00:03:00 2
2013-12-24 00:04:00 2
2013-12-24 00:05:00 1
Note: you may want to handle the edges differently (e.g., 2013-12-24 00:02:00 2)
