ggseas package for seasonal adjustment on the fly with ggplot2

By Peter’s stats stuff – R

In a post a few months ago I built a new ggplot2 statistical transformation (“stat”) to provide X13-SEATS-ARIMA seasonal adjustment on the fly. With certain exploratory workflows this can be useful, saving on a step of seasonally adjusting many different slices and dices of a multi-dimensional time series during data reshaping, and just drawing it for you as part of the graphics definition process.

That code no longer works – one of the hazards of the fast changing data science environment – since the release of ggplot2 version 2.0 just before Christmas. ggplot2 v.2 introduced a whole new system of object oriented programming, ggproto, replacing the old approach that was based on proto. The good news is that the new approach is easy to use and well documented, and it’s been so successful in encouraging development based on ggplot that there’s now a whole website just to keep track of packages extending ggplot.

So it seemed time to wrap my little convenience function for seasonal adjustment into an R package which duly has its home on GitHub. Like my similar posts on this, this work depends on Christophe Sax’s {seasonal} R package, and of course the US Census Bureau’s X13-SEATS-ARIMA software.

X13-SEATS-ARIMA

Here’s how it works, applying SEATS seasonal adjustment with the default outlier etc settings to the classic international Air Passengers example dataset. The original unadjusted data are in pale grey in the background, and the seasonally adjusted line is the dark one:

# installation if not already done
library(devtools)
install_github("ellisp/ggseas/pkg")

library(ggseas)

ggplot(ap_df, aes(x = x, y = y)) +
   geom_line(colour = "grey80") +
   stat_seas(start = c(1949, 1), frequency = 12) +
   ggtitle("SEATS seasonal adjustment - international airline passengers") +
   ylab("International airline passengers per month")

Extra arguments can get passed through to X13 via …read more

Source:: r-bloggers.com