HarlanH

df.by.rows <- function(df, fn, agg.type=numeric, .progress=NULL) 
{
  # A simplified and more-efficient version of aaply for data frames. Iterates over rows of the df, applies the
  # fn to each single-row slice, and aggregates the **scalar** results.
  #
  # Args:
  #    df - data frame to split
  #    fn - function to apply to each row
  #    agg.type - a constructor, one of numeric, character
  #    .progress - set to "text" to get a progress bar

HarlanH

# NOTE: requires ImageMagick to be installed and in the PATH! Should work on most Linux and Mac systems.

library(ggplot2)

d <- data.frame(x=rep(1:10,10), y=rnorm(100, 1:10))

files <- character()

for (i in 1:10)
{

HarlanH

# demonstration of monotonic smoothing with 2-D integral IVs
# Harlan Harris
# [email protected]

library(ggplot2)
library(locfit)
library(monoProc)

df <-
structure(list(counts = c(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 

HarlanH

If you see the bug, keep your damn mouth shut!

Example #1
Show data.
I've written pointless code. 
traceback() shows where the problem lies, but not when or why
Explain "call stack" made of "frames", noting that many things that don't 
syntactically look like function calls in R actually are.
Error doesn't seem to happen all the time, so adding browser() calls 
might not be very helpful.

HarlanH

# solution via Hadley Wickham and others at: 
# http://www.mail-archive.com/[email protected]/msg85203.html

qq <- list(data.frame(c1=c('a', 'b', 'c')), data.frame(c1=c('a','d','e')), data.frame(c1=c('a','f','g')))

fold <- function(x, fun) {
        if (length(x) == 1) return(fun(x))
        
        accumulator <- fun(x[[1]], x[[2]])
        if (length(x) == 2) return(accumulator)

HarlanH

# Demo of techniques to visualize the predictions made by a categorization model.

library(ROCR)
library(ggplot2)

load(url('http://dl.dropbox.com/u/7644953/classifier-visualization.Rdata'))

pred.df$actual.bin <- ifelse(pred.df$actual == 'yes', 1, 0)
pred.df <- pred.df[order(pred.df$predicted, decreasing=TRUE), ]

HarlanH

# Data Science DC Titles Visualization

# Here's how this will work. In a main loop, a parameterized visualization function
# is called every N seconds. Each function gets the source spreadsheet fresh, and 
# generates a visual.

# aspects of this code borrowed from Drew Conway:
# https://raw.github.com/drewconway/ZIA/master/R/better_word_cloud/better_word_cloud.R

library(plyr)

HarlanH

library(XML)
library(RCurl)

# USAGE: ldply(c('9780387962406', '9780387961406', '0387981403'), function(x) ISBNdb(x, 'apikey'))

ISBNdb <- function(isbn, access_key, 
                   isbn.api='http://isbndb.com/api/books.xml?access_key=%s&index1=isbn&value1=%s') {
  isbn <- as.character(isbn)

HarlanH

type _NA
end

macro NA() # for easier typing!
    _NA()
end

NumberData = Union(_NA, Number)
StringData = Union(_NA, String)

HarlanH

vswitch <- function(namedList, default=NA, selector) {
  # Function adapted from Bill Dunlap that implements something along the lines of a vectorized switch statement.
  # http://tolstoy.newcastle.edu.au/R/e8/devel/09/12/1122.html
  #
  # Args:
  #    namedList - e.g., list(times=df$a * df$b, plus=df$a + df$b)
  #    default - a value to assign to elements of selector that aren't matched in namedList
  #    selector - e.g., c('times', 'times', 'plus', 'exp', 'plus')
  #
  # Returns: a vector of values selected from namedList