drewconway

# Function returns quater scores from Wikipedia Super Bown pages
get.scores<-function(numeral) {
    # Base URL for Wikipedia 
    wp.url<-getURL(paste("http://en.wikipedia.org/wiki/Super_Bowl_",numeral,sep=""))
    wp.data<-htmlTreeParse(wp.url, useInternalNodes=TRUE)
    score.html<-getNodeSet(wp.data,"//table[@style='background-color:transparent;']")
    score.table<-readHTMLTable(score.html[[1]])
    score.table<-transform(score.table, SB=numeral)
    return(score.table)
}

drewconway

# A function that converts a given integer into its Roman Numeral equivalent
to.RomanNumeral<-function(x) {
    if(0 < x & x < 5000) {
        x<-as.integer(x) 
        digits<-c(1000,900,500,400,100,90,50,40,10,9,5,4,1)
        numerals<-c("M","CM","D","CD","C","XC","L","XL","X","IX","V","IV","I")
        digits.numerals<-as.data.frame(cbind(digits,numerals), stringsAsFactors=FALSE)
        numeral<-""
        for(i in 1:nrow(digits.numerals)) {
            while(x >= as.numeric(digits.numerals[i,1])) {

drewconway

# File-Name:       twitter_word_cloud.R           
# Date:            2011-01-30                                
# Author:          Drew Conway
# Email:           drew.conway@nyu.edu                                      
# Purpose:         Create a comparative word cloud of two twitter hashtags
# Data Used:       
# Packages Used:   twitteR, tm, ggplot2
# Output File:     Hashtag word cloud
# Data Output:     
# Machine:         Drew Conway's MacBook Pro

drewconway

require(RCurl)
require(XML)

google.counts<-function(s){
    search.url<-paste("http://www.google.com/search?q=",gsub(" ","+",s),sep="")
    search.html<-getURL(search.url)
    parse.search<-htmlTreeParse(search.html,useInternalNodes = TRUE)
    search.nodes<-getNodeSet(parse.search,"//div[@id='resultStats']")
    search.value<-strsplit(xmlValue(search.nodes[[1]])," ",fixed=TRUE)[[1]][2]
    return(as.numeric(gsub(",","",search.value,fixed=TRUE)))

drewconway

def twitter_network(users, api, user_type="search", alt_type="friend"):
    """
    Given a list of Twitter users, create NetworkX object of relationships.
    args:   users           List of Twitter users as strings
            user_types      Type string for entries in 'users'
    """
    twitter_network=nx.DiGraph()
    # Iteratively create network with appropriate type data
    users=list(users)
    for u in users:

drewconway

# Get StackOverflow data
get.stack<-function(tok) {
    # Must check for XML install, thanks onertipaday!
    if (!require(XML)) install.packages('XML')
    library(XML)
    # Enter a SO tag as character string, and number of tags are returned
    tok<-gsub("(/| )","-",tok)
    tok<-gsub("#","%23",tok,fixed=TRUE)
    base.stack<-"http://stackoverflow.com/questions/tagged/"
    stack.tree<-htmlTreeParse(paste(base.stack,tok,sep=""),useInternalNodes=TRUE)

drewconway

#### The following code produces a crash of ggplot2 ####

# Load data
hashtag<-"rstats"
infochimps<-read.csv(paste(hashtag,"_infochimps.csv",sep=""))

# Produce plot
png(paste(hashtag,"_infochimps_metric.png",sep=""),height=800,width=800,res=100)
ic.plot<-ggplot(infochimps,aes(x=log(followers_count/friends_count),y=trstrank))+geom_text(aes(label=screen_name,color=tweet.hash,size=replies_out/replies_in))
ic.plot<-ic.plot+xlab(expression(log[frac(Followers,Friends)]))+ylab("Infochimps.org trstrank")+opts(title=paste("Key Actor Analysis for",hashtag," with Infochimps.org Data",sep=""))

drewconway

### Perform analysis and make pretty pictures ###

png("surv_plot.png",height=1000,width=1000,res=100)
par(mfrow=c(2,2))
survplot(survfit(Surv(time,censor)~type,data=subset(surv_data,type=="lonely")),what="survival",conf="bands",xlab="Minutes")
title("Survival function for seeing a lonely guy")
survplot(survfit(Surv(time,censor)~type,data=subset(surv_data,type=="penis")),what="survival",conf="bands",xlab="Minutes",main="Survival function for seeing a penis")
title("Survival function for seeing a penis")
survplot(survfit(Surv(time,censor)~type,data=subset(surv_data,type=="drunk")),what="survival",conf="bands",xlab="Minutes",main="Survival function for seeing two or more drunk people")
title("Survival function for seeing two or more drunk people")

drewconway

### Generate our data ###

# Time to seeing a penis
penis<-round(rchisq(30,2))
p_cens<-rep(1,30)
p_type<-rep("penis",30)
p_bind<-cbind(penis,p_cens,p_type)

# Time to seeing a lonely dude
lonely<-round(rchisq(40,1))

drewconway

library(survival)
library(ggplot2)
library(Design)


### Test data generation ###

# Chi-square test
count<-1000
cs1<-rchisq(count,1)