I have some ideas on how to do what I want, but I thought I would pose this question as I know that some of the monks may come up with better ideas than mine.

The problem

I have collected some data (blood glucose levels) over about a 4 week period of time. The correlation I would like to do is an average of those readings by time of day. For instance if you look at a subset of the data that I have collected I have:

Blood Glucose by Date Date and Time Blood Glucose Hg/Ml 2004-02-01 07:01 82 2004-02-01 11:38 172 2004-02-01 22:48 154 2004-02-02 05:38 107 2004-02-02 13:20 117 2004-02-02 23:48 188

As you can see by the subset of data my sample times do not occur at exactly the same times every day depending on my schedule, when I get up, what dumb meetings interrupt my flow, etc.

What I want to do is average these readings by time of day such that I calculate averages for from 4:30AM (when I get up days I go to the gym) all the way through until midnight. Some interpolation is going to be called for since I do not test every hour and my samples are essentially 3 to 4 a day (more when I feel crappy).

Thoughts anyone?