aammd

## a demonstration of ordered contrasts

## three levels, A, B, and C
factor1 <- rep(LETTERS[1:3],c(15,15,15))

## Another factor
factor2 <- rep(letters[1:3],15)

## check combinations
table(factor1,factor2)

aammd

Simple data manipulations in R

Many years ago, I was introduced to R by Cam Webb . At the time, his website contained a list of common data manipulations (original here). This list dated from Cam's early experience with R, and contained the R-help mailing list responses to a series of data manipulations. For a long time, I kept this file as a handy reference. I printed it out. I recommended it to friends.

Now I have been using R for years, and the state of the art has advanced considerably. Particulary, Hadley Wickham's reshape2 and dplyr packages have transformed the way most useRs manipulate their data. I decided that it would be interesting to revisit my favourite resource and try my hand at solving these problems with tools from these two packages.

library(reshape2)
library(dplyr)

aammd

Apparently, there is a simple problem called Fizz buzz which is sometimes used to identify competent programmers. A good opportunity to practice some dplyr and magrittr tricks.

library(dplyr)
library(magrittr)
library(knitr)

1:100 %>%
 data.frame %>%

aammd

Is there an easy way to convert a named list into a dataframe, preserving the elements of the list in a "list-column"?

    library(dplyr)
    library(magrittr)

    ## make a random matrix
    rand_mat <- function() {

aammd

set.seed(1)

# Create fake data
x = runif(100, 0, 5)
y = .25 * x^3 - x^2 + rnorm(length(x))
data = data.frame(x = x, y = y)

# Identify 500 points to include in the plots
x.sequence = seq(0, 5, length = 500)

aammd

Bootstrap confidence intervals for linear and nonlinear models

Another version of this gist with figures included is on Rpubs

Recently I was trying to put confidence intervals on a regression line, and I got some excellent advice from @davidjayharris on Twitter, who suggested the below method in an excellent

aammd

On Rosetta Code, the list of Tasks uncompleted in R contains the following challenge: print the Bernoulli numbers from B0 to B60. After seeing some solutions, including one in in Python , I wanted to try in R. There are examples of code on that website, and a nice description of the algorithm here.

 library(MASS)

aammd

## via base ####

"%gcd%" <- function(u, v) {ifelse(u %% v != 0, v %gcd% (u%%v), v)}

n <- 20L
js <- 0:n
Jp1 <- js + 1L
Ns <- matrix(NA,nrow = n + 1, ncol = n + 1)
Ds <- Ns
Ns[1,] <- 1L

aammd

## Empty list -- just use the empty environment for this.
nil <- function() {
  emptyenv()
}

## Test if a list is the empty list:
is_empty <- function(lis) {
  identical(lis, nil())
}

aammd

values <- c(2,5,3,6,7,
            9,5,4,9,9,
            1,5,4,8,1,
            3,1,5,6,2,
            2,9,4,7,4)
my.mat <- matrix(values, nrow = 5, byrow = TRUE)

library(dplyr)
library(tidyr)