drewconway

# Visualize data and save
png("Tax_percentage.png",height=1000,width=1000,res=100)
ggplot(frum_data,aes(tax1))+stat_density()+geom_vline(aes(xintercept=31.5,colour="Actual Value"))+opts(title="In approximate percentage terms, how much is the U.S. (federal)
    government currently taking out of the U.S. economy in taxation?",plot.title=theme_text(size = 12))+xlab("Simulated Tea Party Survey Response Distribution")+ylab("Density")
dev.off()    

png("Tax_family.png",height=1000,width=1000,res=100)
ggplot(frum_data,aes(tax2))+stat_density()+geom_vline(aes(xintercept=7.5,colour="Actual Value"))+opts(title="How much federal income tax do you think a typical family
    earning $50,000 pays (in 1,000 dollars)?",plot.title=theme_text(size = 12))+xlab("Simulated Tea Party Survey Response Distribution")+ylab("Density")
dev.off()

drewconway

# From the FrumForum.com Tea Party survey we see the following result for 
# the questions: 

# Question 1: In approximate percentage terms, how much is the U.S. (federal) 
# government currently taking out of the U.S. economy in taxation?
#
# Mean:     42.06%
# STD:      19.06%
# Actual:   31.5%
# 

drewconway

### Meetup activity history ###

joins<-read.csv("New_York_R_Statistical_Programming_Meetup_Groups_Joins.csv")
rsvp<-read.csv("New_York_R_Statistical_Programming_Meetup_RSVPs.csv")
activity<-read.csv("New_York_R_Statistical_Programming_Meetup_Total_and_Active_Members.csv")

# Merge the data into a single frame
all<-merge(joins,rsvp,by="Date",all=TRUE)
all<-merge(all,activity,by="Date",all=TRUE)
all_dates<-as.vector(all$Date)

drewconway

library(ggplot2)
library(XML)

### Meetup topics word cloud ###

# Get the raw meetup description into a dataframe
raw_desc<-levels(read.table('descriptions.txt',sep="\n")$V1)
clean_strings<-function(s){
    low<-tolower(s)
    clean<-gsub("[[:punct:]\n]","",low)

drewconway

library(geoR)
library(geoRglm)

# Perfrom MCMC simulations
model<-list(cov.pars = c(1, 1), beta = 1, family = "poisson")
mcmc.test<-mcmc.control(S.scale = 0.45, thin = 1)
test.tune<-glsm.mcmc(haiti.geo, model = model, mcmc.input=mcmc.test)
haiti.mcmc<-prepare.likfit.glsm(test.tune)
prior<- prior.glm.control(phi.prior = "fixed", phi = .1)

drewconway

# File-Name:       Scotch_Pref.R                 
# Date:            2009-11-29                                
# Author:          Drew Conway                                       
# Purpose:         Display one-dimensional item response for scotch whiskey preference
# Data Used:       whiskey, package=flexmix
# Packages Used:   zelig,ggplot
# Output File:     scotch_pref.png
# Data Output:     
# Machine:         Drew Conway's MacBook 

drewconway

# File-Name:       cpi_oprobit.R                 
# Date:            2009-11-17                                
# Author:          Drew Conway                                       
# Purpose:         Quick ordered probit analysis of the Corruption Perceptions Index 2009
#                   to check for effect of number of surveys used on CPI scores
# Data Used:       corruption_index.csv
#  available here: http://www.drewconway.com/zia/wp-content/uploads/2009/11/corruption_index.csv
# Packages Used:   foreign,Zelig
# Output File:     
# Data Output:     

drewconway

# File-Name:       currency_converter.R                 
# Date:            2009-11-17                                
# Author:          Drew Conway                                       
# Purpose:         Convert currency data
# Data Used:       vc_invests.csv
# Packages Used:   foreign,XML
# Output File:     vc_invests_USD.csv
# Data Output:     
# Machine:         Drew Conway's MacBook                         
                                                                    

drewconway

def write_data(data,path,new_path):
# Takes data dict and writes new data to a new file
    reader=csv.reader(open(path,'U'),delimiter=',')
    writer=csv.writer(open(new_path,"w"))
    row_num=0
    for row in reader:
        if row_num<1:
        # Keep ther same column headers as before, so we simply
        # re-write the first row.
            writer.writerow(row)

drewconway

import html5lib
from html5lib import treebuilder

def parse_data(player_urls):
# Returns a dict of player data parse trees indexed by player name
    # Create a dict indexed by player names
    player_data=dict.fromkeys(player_urls.keys())
    # Download player profile data and parse using html5lib
    for name in player_urls.keys():
    # html5lib integrates the easy-to-use BeautifulSoup parse tree using the treebuilders library.