Saturday, July 16, 2011

Histogram in Terminal

I'm sure there's an R packages with this kind of thing out there, but I didn't find it.  Often when working with MCMC on a server over ssh, you need to get a rough idea about parameter distributions and convergence.  This code outputs a reasonable histogram on the command line without needing graphics.  Something similar for traces would be nice.



    histo <- function(x,...) {
        o <- hist(x, plot = FALSE, ...)
        w <- options()[['width']]
        labels <- signif(o[['mids']],2)
        pad <- max(nchar(labels))
        densities <- o[['density']]/max(o[['density']]) * (w-pad)
        for ( i in 1:length(densities)) {
            padI <- pad - nchar(labels[i]) + 1
            if ( labels[i] >= 0 ) {
                cat(' ', labels[i], rep(' ',padI-1), '|', sep='')
            } else {
                cat(labels[i], rep(' ',padI  ), '|', sep='')
            }
            cat(rep('*',densities[i]), sep='')
            cat('\n', sep='')
        }
        return(o)
    }

    h <- histo(tp, breaks = 30)

The height (horizontal) of the histogram bars comes from the options, so if it's printing too wide for you, set "options(width = ???)" to something that works.

1 comment:

  1. You know what would make it all better? A flowchart.

    ReplyDelete