Plot each year of a time series on the same x-axis using Pandas
I wanted to compare several years of daily albedo observations to one another by plotting them on the same x (time) axis. You can do this by taking advantage of Pandas’ pivot table functionality.
The data are contained in a pandas Series, indexed with datetimes:
al_pd.head()
TIME
2010-05-01 73.538643
2010-05-02 65.370290
2010-05-03 69.298476
2010-05-04 70.018208
2010-05-05 76.632503
...
dtype: float64
First smooth the time series by taking the mean over a rolling triangular window of width 3 values:
al_pd = al_pd.rolling(window=3, center=True, win_type='triang').mean()
Convert to a DataFrame and add two columns, one containing the day-of-year and the other the year:
al_pd.name = 'Albedo'
al_pd = al_pd.to_frame()
al_pd['doy'] = al_pd.index.dayofyear
al_pd['Year'] = al_pd.index.year
The dataframe now looks like this:
Albedo doy Year
TIME
2010-05-01 73.538643 121 2010
2010-05-02 65.370290 122 2010
2010-05-03 69.298476 123 2010
2010-05-04 70.018208 124 2010
2010-05-05 76.632503 125 2010
...
Finally create a pivot table, which will have a column for each year in the data set and a row for each day-of-year:
piv = pd.pivot_table(al_pd, index=['doy'],columns=['Year'], values=['Albedo'])
This outputs a pivot table which looks like this:
Albedo
Year 2010 2011 2012 2013 2014 2015
doy
121 73.538643 71.156460 70.771080 69.188382 71.635948 78.025509
122 65.370290 68.695741 75.071027 74.486435 68.093146 75.218588
123 69.298476 73.066383 69.339166 74.673265 65.952841 75.665903
124 70.018208 69.254242 69.014770 77.758446 63.181904 70.147964
125 76.632503 71.679306 68.095796 72.417934 72.567538 74.658602
...
The pivot table is simple to plot:
import seaborn as sns
piv.plot()
Written on June 21, 2016