forked from pandas-dev/pandas
-
Notifications
You must be signed in to change notification settings - Fork 27
Time Series Manipulation
adamklein edited this page Mar 15, 2012
·
11 revisions
It has been agreed upon on the Mailing List, that functionality of the TimeSeries Scikit into Pandas. This page shall list
- workflows in different fields of science
- required functionality
- link to already developed code as input
Maybe that needed functionality already exists with pandas. In this case, useful examples are needed.
Please list your workflows below in order to enable fellow developers to consider your use case.
The following functions from the scikit shall be ported to pandas:
- Loading a TimeSeries from a text file for conveniently read a timeseries from a ASCII file.
- convert_to_annual. E.g. for easily taking annual averages considering leap days.
- Easy save time series to HDF
- Plotting with automatic tick and label formatting
- Reporting that considers lag, offset, nodata, date formatting, leap year, aggregated years.
Following functionality could be added for convenient timeseries manipulation:
- support for aggregated series: e.g. long-term averages of hourly data
- all the frequency constants
- extras.py
- convert_to_annual
- guess_freq
- tsfromtxt (ts file loading, saving)
- parser.py
- incorporate in addition to default dateutil parser
- tseries.py
- has_missing_dates
- is_full (opposite of has_missing dates)
- pct, pct_log, pct_symmetric
- lib/plotlib.py
- fancy plotting
additional notes: need to expose strftime on interval to user code