R

What is R

R

• the eighteenth letter of the alphabet

• the letter before S

• a language and environment for statistical computing and graphics

  • linear and nonlinear modeling, classical statistical tests, time-series analysis, classification, clustering
  • scatterplots, time series, histogram, bar plot, pie chart, image plot

• R Reference Card

R (continued)

• The R Ecosystem

  • over 2 million users
  • thousands of package developers

• simple procedural programming language with functions

  • also has object-oriented programming facilities

• extensible through user-submitted packages

• C, C++, and Fortran code can be linked in and called

• C and Java code can call into R

Why R

• your Portfolio Manager procastinated on this week’s topic

• interest in combining R’s strength in data manipulation, statistics, and data visualization with big data in Hadoop

• there’s clues about how Google and Facebook use R

  • Google R Style Guide
  • Predicting the Present with Google Trends (Google Research)
  • What’s on your mind? (Facebook)

• Revolution Analytics produces a product for connecting R to Hadoop

• also available as R packages on GitHub

R Language

Variables

	> x = 3.14     # assign 3.14 to x
	> x            # print the value of x
	[1] 3.14

Numerics, integers

	> x = 3.14
	> class(x)     # print the class name of x
	[1] "numeric"  # numer is essentially “double”

	> y = as.integer(3.14)
	> y
	[1] 3
	> class(y)
	[1] "integer"
	> is.integer(y)
	[1] TRUE

Logical

	> t = TRUE
	> f = FALSE
	> class(t)
	[1] "logical"
	> t
	[1] TRUE
	> !t
	[1] FALSE
	> t & f      # t AND f
	[1] FALSE
	> t | f      # t OR f
	[2] TRUE

Character

	> n = 'Nancy'
	> x = as.character(3.14)
	> class(x)
	[1] "character"
	> paste(n, x)      # combine strings
	[1] "Nancy 3.14"
	> sprintf("%s has %d dollars", n, as.double(x))
	[1] "Nancy has 3.14 dollars"

Vector

• an ordered collection of the same type

  > c(1, 2, 3)
  [1] 1 2 3
  > c("a", "bb", "ccc", "dddd")
  [1] "a" "bb" "ccc" "dddd"
  > c(1, "a", TRUE)
  [1] "1" "a" "TRUE"
  > a = c(1, 2, 3)
  > b = c(10, 20, 30, 40, 50, 60)
  > a + b
  [1] 11 22 33 41 52 63

List

• an ordered collection of objects

  > a = c(1, 2, 3)
  > l = list(a, TRUE, "ccc")
  > l
  [[1]]
  [1] 1 2 3
  [[2]]
  [1] TRUE
  [[3]]
  [1] "ccc"

Frame

• a data table

  > p = c(1.99, 2.99, 4.99)
  > c = c(51, 24, 36)
  > data.frame(p, c)
       p  c
  1 1.99 51
  2 2.99 24
  3 4.99 36
  > df = data.frame(price=p, count=c)
  > rownames(df) = c("squash", "cucumber", "tomato")

Time for Action

Load Data

• CSV > pt = read.csv(‘Parking_Tags_Data_2012-10000.csv’)
• Microsoft Excel > install.packages() # install the gdata package if needed > library(gdata) > pt = read.xls(‘Parking_Tags_Data_2012-10000.xls’)
• use head() to peek at the first few rows of a data frame

Scatter Plot

• basic plot > plot(pt$infraction_code, pt$set_fine_amount) > plot(pt$infraction_code, pt$set_fine_amount, xlab=”Infraction Code”, ylab=”Fine Amount”)

Histogram

• frequency distribution > hist(pt$set_fine_amount)

Facebook Friends

• kinghuang/facebook_mining.r

• kinghuang/facebook_friendship1.r

• rakosnicek/facebook_friendship2.r

• visit Graph API Explorer to generate an Access Token

  > source('facebook_mining.r')
  > source('facebook_friendship1.r')

Twitter

• Get an OAuth token
• to authenticate library(twitteR) cred = OAuthFactory$new(consumerKey=”“, consumerSecret=”“, requestURL=”https://api.twitter.com/oauth/request_token”, accessURL=”https://api.twitter.com/oauth/access_token”, authURL=”https://api.twitter.com/oauth/authorize”) cred$handshake() registerTwitterOAuth(cred)
• get tweets userTimeline(“kinghuang”, n=20)
• see Text Mining

Exercise

• take some data and use R to derive something

  • frequent words (Twitter example)

  • relationships between items (Facebook Example)

  • R-bloggers articles

    • Do the Simpsons characters like each other?
    • ggplot2 Chloropleth of Supreme Court Decisions
    • Rain in Netherlands during the past 100 years

• possible data sources

  • Calgary, Edmonton, and Toronto Open Data
  • Twitter, Facebook
  • StackExchange Data Explorer
  • Best Buy BBYOPEN