[SciPy-user] Calculating daily, monthly and seasonal averages of hourly time series data.

Pierre GM pgmdevlist@gmail....
Thu Oct 9 12:17:28 CDT 2008


Dharhas,
What you need for timeseries is a recent numpy (>=1.2.0) and 
scipy(>=0.7svn...). If you can compile the latest sources from SVN, you're 
good to go.

The documentation is here:
http://pytseries.sourceforge.net/

If it doesn't cover areas you need, let us know and help us by writing some 
examples/tutorial/whatever.

As an extra comment on Lionel's answers:
* Once you have a time series, you can directly mask some data without having 
to recreate a series
>>> series=ts.time_series(np.random.rand(365), start_date=ts.now('D'))
>>># Mask the first 10 elements
>>> series[:10] = ma.masked
>>> #Maske the data in August
>>> series[series.month==8] = ma.masked


More information about the SciPy-user mailing list