Showing posts with label time series. Show all posts
Showing posts with label time series. Show all posts

Monday, December 9, 2013

Data mountains and streams - stacked area plots in R



Below are two functions for producing stacked area plots. The first is the more typical approach where sequential series are stacked on top of another (function: plot.stacked), while the second approach is the more aesthetically-oriented version called a "stream plot" (function: plot.stream), which alternates series on either side of a meandering baseline (see here for the motivation, and here for the inspiration). 

Arguments are similar for both functions regarding the input of x and y series and polygon attributes (fill color, border color, border line width). The stream plot also requires that the degree of meandering for the baseline be defined by the arguments frac.rand and spar; frac.rand, controls the meander amplitude (uniform random numbers added to baseline as a fraction of the total y range) and spar controls the amount of smoothing (as fit by the function smooth.spline).

The plot above colors the series with a color gradient of when the first appear in the series, while the plot below colors series by their maximum value. The order of the plotting of the series can also affect the the emphasis on the plot. By default, plotting order is sequential by column, although two ordering options are built-in to the functions: order by maximum value, and order by first appearance.



The plot.stacked function:

Thursday, January 10, 2013

Lomb-Scargle periodogram for unevenly sampled time series


In the natural sciences, it is common to have incomplete or unevenly sampled time series for a given variable. Determining cycles in such series is not directly possible with methods such as Fast Fourier Transform (FFT) and may require some degree of interpolation to fill in gaps. An alternative is the Lomb-Scargle method (or least-squares spectral analysis, LSSA), which estimates a frequency spectrum based on a least squares fit of sinusoid.

The above figure shows a Lomb-Scargle periodogram of a time series of sunspot activity (1749-1997) with 50% of monthly values missing. As expected (link1, link2), the periodogram displays a a highly significant maximum peak at a frequency of ~11 years.

The function comes from a nice set of functions that I found here: http://research.stowers-institute.org/efg/2005/LombScargle/R/index.htm. An accompanying paper focusing on its application to time series of gene expression can be found here.

Below is a comparison to an FFT of the full time series. For another great resource on spectral analysis, and time series-related R methods in general, see the following website: http://zoonek2.free.fr/UNIX/48_R/15.html.


To reproduce the example: