[SciPy-user] time series analysis
Matt Knox
mattknox_ca at hotmail.com
Tue Nov 6 06:38:17 EST 2007
> Some things I can imagine are the following:
> create a tools directory under the timeseries tree.
There is a "lib" sub-directory for stuff that falls outside the core
Date/TimeSeries classes. It currently includes a sub-module for "moving
functions" (moving average, etc...), and interpolation.
> * common frequency conversions: reduce to hourly values
Frequency conversions are simple to do using the "convert" method of the
TimeSeries class. Here is an example converting an hourly frequency series to
daily...
>>> import numpy as np
>>> import maskedarray as ma
>>> import timeseries as ts
>>> h = ts.time_series(np.arange(50, dtype=np.float32), start_date=ts.today
('hourly'))
>>> h
timeseries([ 0. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14.
15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.
30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.
45. 46. 47. 48. 49.],
dates = [06-Nov-2007 06:00 ... 08-Nov-2007 07:00],
freq = H)
>>> d = h.convert('daily')
>>> d
timeseries(
[[-- -- -- -- -- -- 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
13.0 14.0 15.0 16.0 17.0]
[18.0 19.0 20.0 21.0 22.0 23.0 24.0 25.0 26.0 27.0 28.0 29.0 30.0 31.0
32.0 33.0 34.0 35.0 36.0 37.0 38.0 39.0 40.0 41.0]
[42.0 43.0 44.0 45.0 46.0 47.0 48.0 49.0 -- -- -- -- -- -- -- -- -- -- --
-- -- -- -- --]],
dates =
[06-Nov-2007 ... 08-Nov-2007],
freq = D)
>>> d_avg = h.convert('daily', ma.average)
>>> d_avg
timeseries([ 8.5 29.5 45.5],
dates = [06-Nov-2007 ... 08-Nov-2007],
freq = D)
===============================================
If any of the above seems mysterious, let me know and I can offer a more
detailed explanation.
> * error checking of measurement data: statistically and logically
Some data error checking algorithms could be useful, yes. I won't likely be
working on them in the near future though.
> Is it safe for me to replace/patch my current maskarray?
Generally speaking, the maskedarray package is mostly backwards compatible
with the current numpy.ma package , and thus the api is very stable so you
should be able to update it without any problems.
> I have experience in compiling on linux but have
> to work on a windows box.
> Therefore an advice on that would be useful.
If you are using Python 2.5 on windows, it is easy to do using mingw. If you
are using an earlier version of Python, Visual studio 2003 is the easiest way
to go.
> Is there a possibility to subscibe to SVN to get an email on chnages?
I don't know of any svn clients that have a built in way to do this, but it
likely wouldn't be difficult to write your own script to do this since
subversion has a command line interface.
- Matt
More information about the SciPy-User
mailing list