Created
September 9, 2013 19:05
-
-
Save renecnielsen/6500025 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
### Get, save and load data ### ----- | |
# temp holder file | |
# create download directory and set it | |
.exdir = '~/dir' | |
dir.create(.exdir) | |
.file = file.path(.exdir, 'MYWorld_votes_all.csv.tar.bz2') | |
# Download file | |
url = 'http://ec2-23-22-13-62.compute-1.amazonaws.com/dataset/data/MYWorld_votes_all.csv.tar.bz2' | |
download.file(url, .file) | |
# Untar it | |
untar(.file, compressed = 'bzip2', exdir = path.expand(.exdir)) | |
# Set Working Directory and set dataset | |
setwd("~/dir/") | |
myworld <- read.csv("~/dir/MYWorld_votes_all.csv") | |
# View(myworld) | |
### Labels ### ----- | |
# Read in labels. | |
require(RCurl) | |
countryLabelURL <- getURL("https://docs.google.com/spreadsheet/pub?key=0Agx2NHe4DFa9dFNxUnh4Q204TnNKVFRpcl9obEJSdmc&single=true&gid=1&output=csv") | |
countryLabels <- read.csv(textConnection(countryLabelURL)) | |
genderLabelURL <- getURL("https://docs.google.com/spreadsheet/pub?key=0Agx2NHe4DFa9dFNxUnh4Q204TnNKVFRpcl9obEJSdmc&single=true&gid=4&output=csv") | |
genderLabels <- read.csv(textConnection(genderLabelURL)) | |
educationLabelURL <- getURL("https://docs.google.com/spreadsheet/pub?key=0Agx2NHe4DFa9dFNxUnh4Q204TnNKVFRpcl9obEJSdmc&single=true&gid=2&output=csv") | |
educationLabels <- read.csv(textConnection(educationLabelURL)) | |
prioritiesLabelURL <- getURL("https://docs.google.com/spreadsheet/pub?key=0Agx2NHe4DFa9dFNxUnh4Q204TnNKVFRpcl9obEJSdmc&single=true&gid=3&output=csv") | |
prioritiesLabels <- read.csv(textConnection(prioritiesLabelURL)) | |
# Add labels | |
myworld$country <- countryLabels$label[match(myworld$country, countryLabels$id)] | |
myworld$gender <- genderLabels$label[match(myworld$gender, genderLabels$id)] | |
myworld$education <- educationLabels$label[match(myworld$education, educationLabels$id)] | |
myworld$priority1 <- prioritiesLabels$label[match(myworld$priority1, prioritiesLabels$id)] | |
myworld$priority2 <- prioritiesLabels$label[match(myworld$priority2, prioritiesLabels$id)] | |
myworld$priority3 <- prioritiesLabels$label[match(myworld$priority3, prioritiesLabels$id)] | |
myworld$priority4 <- prioritiesLabels$label[match(myworld$priority4, prioritiesLabels$id)] | |
myworld$priority5 <- prioritiesLabels$label[match(myworld$priority5, prioritiesLabels$id)] | |
myworld$priority6 <- prioritiesLabels$label[match(myworld$priority6, prioritiesLabels$id)] | |
### Quick Overview ### ----- | |
head(myworld) | |
summary(myworld) | |
### Creating a new variable for each priority ### ----- | |
myworld$climate <- with(myworld, ifelse(priority1 == "Action taken on climate change" | | |
priority2 == "Action taken on climate change" | | |
priority3 == "Action taken on climate change" | | |
priority4 == "Action taken on climate change" | | |
priority5 == "Action taken on climate change" | | |
priority6 == "Action taken on climate change", 1, 0)) | |
table(myworld$climate) | |
myworld$transport <- with(myworld, ifelse(priority1 == "Better transport and roads" | | |
priority2 == "Better transport and roads" | | |
priority3 == "Better transport and roads" | | |
priority4 == "Better transport and roads" | | |
priority5 == "Better transport and roads" | | |
priority6 == "Better transport and roads", 1, 0)) | |
table(myworld$transport) | |
myworld$support <- with(myworld, ifelse(priority1 == "Support for people who can't work" | | |
priority2 == "Support for people who can't work" | | |
priority3 == "Support for people who can't work" | | |
priority4 == "Support for people who can't work" | | |
priority5 == "Support for people who can't work" | | |
priority6 == "Support for people who can't work", 1, 0)) | |
table(myworld$support) | |
myworld$water <- with(myworld, ifelse(priority1 == "Access to clean water and sanitation" | | |
priority2 == "Access to clean water and sanitation" | | |
priority3 == "Access to clean water and sanitation" | | |
priority4 == "Access to clean water and sanitation" | | |
priority5 == "Access to clean water and sanitation" | | |
priority6 == "Access to clean water and sanitation", 1, 0)) | |
table(myworld$water) | |
myworld$health <- with(myworld, ifelse(priority1 == "Better healthcare" | | |
priority2 == "Better healthcare" | | |
priority3 == "Better healthcare" | | |
priority4 == "Better healthcare" | | |
priority5 == "Better healthcare" | | |
priority6 == "Better healthcare", 1, 0)) | |
table(myworld$health) | |
myworld$edu <- with(myworld, ifelse(priority1 == "A good education" | | |
priority2 == "A good education" | | |
priority3 == "A good education" | | |
priority4 == "A good education" | | |
priority5 == "A good education" | | |
priority6 == "A good education", 1, 0)) | |
table(myworld$edu) | |
myworld$government <- with(myworld, ifelse(priority1 == "A responsive government we can trust" | | |
priority2 == "A responsive government we can trust" | | |
priority3 == "A responsive government we can trust" | | |
priority4 == "A responsive government we can trust" | | |
priority5 == "A responsive government we can trust" | | |
priority6 == "A responsive government we can trust", 1, 0)) | |
table(myworld$government) | |
myworld$phone <- with(myworld, ifelse(priority1 == "Phone and internet access" | | |
priority2 == "Phone and internet access" | | |
priority3 == "Phone and internet access" | | |
priority4 == "Phone and internet access" | | |
priority5 == "Phone and internet access" | | |
priority6 == "Phone and internet access", 1, 0)) | |
table(myworld$phone) | |
myworld$energy <- with(myworld, ifelse(priority1 == "Reliable energy at home" | | |
priority2 == "Reliable energy at home" | | |
priority3 == "Reliable energy at home" | | |
priority4 == "Reliable energy at home" | | |
priority5 == "Reliable energy at home" | | |
priority6 == "Reliable energy at home", 1, 0)) | |
table(myworld$energy) | |
myworld$food <- with(myworld, ifelse(priority1 == "Affordable and nutritious food" | | |
priority2 == "Affordable and nutritious food" | | |
priority3 == "Affordable and nutritious food" | | |
priority4 == "Affordable and nutritious food" | | |
priority5 == "Affordable and nutritious food" | | |
priority6 == "Affordable and nutritious food", 1, 0)) | |
table(myworld$food) | |
myworld$forests <- with(myworld, ifelse(priority1 == "Protecting forests, rivers and oceans" | | |
priority2 == "Protecting forests, rivers and oceans" | | |
priority3 == "Protecting forests, rivers and oceans" | | |
priority4 == "Protecting forests, rivers and oceans" | | |
priority5 == "Protecting forests, rivers and oceans" | | |
priority6 == "Protecting forests, rivers and oceans", 1, 0)) | |
table(myworld$forests) | |
myworld$politicalfreedoms <- with(myworld, ifelse(priority1 == "Political freedoms" | | |
priority2 == "Political freedoms" | | |
priority3 == "Political freedoms" | | |
priority4 == "Political freedoms" | | |
priority5 == "Political freedoms" | | |
priority6 == "Political freedoms", 1, 0)) | |
table(myworld$politicalfreedoms) | |
myworld$crime <- with(myworld, ifelse(priority1 == "Protection against crime and violence" | | |
priority2 == "Protection against crime and violence" | | |
priority3 == "Protection against crime and violence" | | |
priority4 == "Protection against crime and violence" | | |
priority5 == "Protection against crime and violence" | | |
priority6 == "Protection against crime and violence", 1, 0)) | |
table(myworld$crime) | |
myworld$discrimination <- with(myworld, ifelse(priority1 == "Freedom from discrimination and persecution" | | |
priority2 == "Freedom from discrimination and persecution" | | |
priority3 == "Freedom from discrimination and persecution" | | |
priority4 == "Freedom from discrimination and persecution" | | |
priority5 == "Freedom from discrimination and persecution" | | |
priority6 == "Freedom from discrimination and persecution", 1, 0)) | |
table(myworld$discrimination) | |
myworld$equality <- with(myworld, ifelse(priority1 == "Equality between men and women" | | |
priority2 == "Equality between men and women" | | |
priority3 == "Equality between men and women" | | |
priority4 == "Equality between men and women" | | |
priority5 == "Equality between men and women" | | |
priority6 == "Equality between men and women", 1, 0)) | |
table(myworld$equality) | |
myworld$job <- with(myworld, ifelse(priority1 == "Better job opportunities" | | |
priority2 == "Better job opportunities" | | |
priority3 == "Better job opportunities" | | |
priority4 == "Better job opportunities" | | |
priority5 == "Better job opportunities" | | |
priority6 == "Better job opportunities", 1, 0)) | |
table(myworld$job) | |
### Save new file as dataframe and .csv ### ----- | |
save(myworld, file="myworld.Rda") | |
write.csv(myworld, file="myworld.csv") |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment