[Tcl-nap-users] time series analysis

[<ni...@ma...>]

Hello,

Have a colleague that would like to analyse some soil temperature
readings, take hourly over the course of 5 years, in 18 different
locations.

Our first attempt was to use the mean_average function, with a 24 window
and 1 hour step.

I have a few questions here...

I am able to get the moving average if I take each column (vector) from my
"data" matrix as

nap "station_one = data(, 0)"
nap "station_one_avg = moving_average(station_one, 24, 1)"

1. how do do this op on the whole matrix, data. I cant seem to get the
example in the docs to work for me. I'll keep playing with it...

2. We really would like to use a moving std. deviation in stead of average.
Is there such a function available, or some help on where to modify the
moving_average proc to yield a moving deviation.

My colleague says this would help him predict/indicate when there was snow
on the ground. Any better approach here???

regards,

nikos