Irregularly-sampled time series data

I have irregularly-sampled time series data. I would like to regularize the data so that every row corresponds to a fixed time period, with the irregular samples summed/divided to match each time bucket. I’m not seeing how to do this with the Time Series tools.

This seems like two tasks: binning and averaging. I don't want to have to give separate time values, but say 'every 5 second' as the bins.

Given:

01:01:01,2

01:01:04,2

01:01:07:3

01:01:08,6

I would like to get this, summing the previous values but averaged across the new time barrier:

01:01:00,0

01:01:05,2+(3*1/3s)=2.33

01:01:10,4+(3*2/3s)=4.66

The sample at :07 had to be averaged across the :05 boundary.

Hi Jack,

 

I'm just thinking about this case and, sorry, but I don't get it.

 

For example in the 01:01:05 timepoint (or is it a timeslot from 0 to 5?)

you say 2 + 3*1/3. Where is the 2 coming from? from time point 1 or time point 4

And the 1/3 comes from that we are looking at 5 don't know 5 (but 4) so 5-4 gets 1. And the 3 from 7-4 = 3?

 

And at timepoint 10, why don't we use the 6 from timepoint 8?

Iris