survey.r

survey.r

### Packages

install.packages(c( "rstudioapi", "corrplot"))

suppressWarnings(
  suppressMessages({
    library(rstudioapi)
    library(readr)
    library(psych)
    library(corrplot)
  })
)

### Setup
current_path = getActiveDocumentContext()$path 
setwd(dirname(current_path ))
print( getwd() )

# setup PDF output (multiple page)
pdf('survey.plot.pdf', paper='a4')

# Load data in a dataframe
dataset = read_delim("survey.data.csv", ";", locale = locale(), trim_ws = TRUE)

# Questions, used for graph title
questions = c(
  firstheard = 'When did you first hear about Bitcoin?',
  DateInvolved = 'Since when are you involved in Bitcoin?',
  involvedStill = 'Are you still involved in Bitcoin?',
  stopBeeingInvolved = 'When did you stop being involved in bitcoin?',
  involvedBecause_SQ1 = 'You got invovled in bitcoin because of its techical aspects.',
  involvedBecause_SQ2 = 'You got invovled in bitcoin because of its political aspect.',
  involvedBecause_SQ3 = 'You got invovled in bitcoin to make money.',
  involvedBecause_SQ5 = 'You got invovled in bitcoin because you share the ideology of the project',
  todayInvolvedReason_SQ1 = 'Today you are involved in bitcoin because bitcoin can change the society',
  todayInvolvedReason_SQ2 = 'Today you are involved in bitcoin because you are interested in the bitcoin technology',
  todayInvolvedReason_SQ3 = 'Today you are involved in bitcoin because it\'s your job or part of your job.',
  todayInvolvedReason_SQ4 = 'Today you are involved in bitcoin because you are investing in Bitcoin',
  todayInvolvedReason_SQ5 = 'Today you are involved in bitcoin because you are mining a significant amount of bitcoin',
  timeInvestedThisYear = 'How many hours of your personal time did you spend per week the last 6 months in the bitcoin project (In average)?',
  motivationComment = 'Any comments about this section?',
  libauth_li1 = 'A woman has to have children in order to be fulfilled|A woman can be fulfilled through her professional career',
  libauth_li2 = 'A woman who does not want to have a child should be allowed to have a free and safe abortion|Abortion should not be allowed in any case',
  libauth_li3 = 'Children should be taught to obey authority |Children should be encouraged to have an independent judgment',
  libauth_li4 = 'People who break the law should get stiffer sentences|Stiffer sentences do not contribute to reducing criminality',
  libauth_li5 = 'Homosexual couples should be able to adopt children|Homosexual couples should not be allowed to adopt children under any circumstances',
  leftright_lr1 = 'Incomes should be made more equal|We need larger income differences as incentives',
  leftright_lr2 = 'People should take more responsibility to provide for themselves|The government should take more responsibility to ensure that everyone is provided for',
  leftright_lr3 = 'People who are unemployed should have to take any job available or lose their unemployment benefits|People who are unemployed should have the right to refuse a job they do not want',
  leftright_lr4 = 'Competition is good. It stimulates people to work hard and develop new ideas|Competition is harmful. It brings the worst in people',
  leftright_lr5 = 'The government should decrease taxes significantly and spend much less on social benefits and services|The government should increase taxes significantly and spend much more on social benefits and services',
  importantActors_SQ1 = 'In your opinion, what are the most important/influential individuals involved in bitcoin (full name or pseudo)? [Most important]',
  importantActors_SQ4 = 'In your opinion, what are the most important/influential individuals involved in bitcoin (full name or pseudo)? [Second important]',
  importantActors_SQ1 = 'In your opinion, what are the most important/influential individuals involved in bitcoin (full name or pseudo)? [Third important]',
  importantActors_SQ1 = 'In your opinion, what are the most important/influential individuals involved in bitcoin (full name or pseudo)? [Fourth important]',
  culture_SQ1 = 'You are familiar with cyberpunk culture',
  culture_SQ2 = 'You are familiar with cypherpunk culture',
  culture_SQ3 = 'The book \'neuromancer\' is an important book',
  belief_SQ1  = 'We need a decentralized currency/asset.',
  belief_SQ2  = 'We cannot trust central banks.',
  belief_SQ3  = 'We cannot trust governments.',
  belief_SQ4  = 'Cryptocurrencies make central banks obsolete.',
  belief_SQ5  = 'Cryptocurrencies are a solution to avoid inflation.',
  belief_SQ6  = 'Privacy protection is an important feature of cryptocurrency.',
  belief_SQ7  = 'Where access to financial institutions is lacking, cryptocurrency will field the gap.',
  belief_SQ8  = 'Bitcoin is a digital gold.',
  belief_SQ9  = 'The government of your country should have the right to keep people under video surveillance in public area.',
  belief_SQ10  = 'The government of your country should have the right to monitor e-mails and any other information exchanged on the internet.',
  belief_SQ11  = 'All government information should be publicly available, even if this means a risk to public security.',
  gender  = 'Your gender',
  country  = 'You live in',
  age = 'Your age',
  agreeToBeContacted = 'Interested to talk a bit more about this on chat or a call?',
  feedback = 'Do you want to add a comment or feedback?'
  )


### Factor some cols
dataset$country = factor(dataset$country, ordered = FALSE )
dataset$gender = factor(dataset$gender, levels = c('Male', 'Female'), ordered = FALSE )
dataset$agreeToBeContacted = factor(dataset$agreeToBeContacted, levels = c('Yes', 'No'), ordered = FALSE )
dataset$involvedStill = factor(dataset$involvedStill, levels = c('Yes', 'No'), ordered = FALSE )

### Factor likert scale
likertLevel = c("Strongly disagree", "Disagree", "Neither agree nor disagree", "Agree", "Strongly agree")
likertCols = c(
  'involvedBecause_SQ1',
  'involvedBecause_SQ2',
  'involvedBecause_SQ3',
  'involvedBecause_SQ5',
  'todayInvolvedReason_SQ1',
  'todayInvolvedReason_SQ2',
  'todayInvolvedReason_SQ3',
  'todayInvolvedReason_SQ4',
  'todayInvolvedReason_SQ5',
  'culture_SQ1',
  'culture_SQ2',
  'culture_SQ3',
  'belief_SQ1',
  'belief_SQ2',
  'belief_SQ3',
  'belief_SQ4',
  'belief_SQ5',
  'belief_SQ6',
  'belief_SQ7',
  'belief_SQ8',
  'belief_SQ9',
  'belief_SQ10',
  'belief_SQ11'
)


dataset[likertCols] = lapply(dataset[likertCols], function(x) factor(x, levels = likertLevel, ordered = TRUE ))

## Check structure
str(dataset, give.attr=FALSE)
sapply(dataset, class)



### Correlation
#correlations = cor(dataset) # ERROR (must be numeric)!
#cor.test(dataset[,todayInvolvedReason_SQ4], dataset[,todayInvolvedReason_SQ5]) #ERROR WTF!
#corrplot(correlations, method="color", addCoef.col="gray")

### Plot all likert
for(likertCol in likertCols) {
  likertCol
  barplot(prop.table(table(dataset[,likertCol])) * 100, col=gray(0:5 / 5), main = paste(strwrap(paste(likertCol, ':', questions[likertCol]), width = 50), collapse = "\n") ,ylim=c(0,100),ylab="%")
}


########

plot(dataset$gender, main = questions['gender'])
plot(dataset$country, main = questions['country'])
plot(dataset$agreeToBeContacted, main = questions['agreeToBeContacted'])
barplot(prop.table(table(format(as.Date(dataset$DateInvolved, format="%d/%m/%Y"),"%Y"))), main = questions['DateInvolved'])
barplot(prop.table(table(format(as.Date(dataset$firstheard, format="%d/%m/%Y"),"%Y"))), main = questions['firstheard'])
barplot(prop.table(table(format(as.Date(dataset$stopBeeingInvolved, format="%d/%m/%Y"),"%Y"))), main = questions['stopBeeingInvolved'])
plot(dataset$involvedStill, main = questions['involvedStill'])

boxplot(dataset$libauth_li1,
        dataset$libauth_li2, # inverted
        dataset$libauth_li3,
        dataset$libauth_li4,
        dataset$libauth_li5  # inverted
        )
title("Libauth")

boxplot(dataset$leftright_lr1,
        dataset$leftright_lr2, #right left
        dataset$leftright_lr3, #right left
        dataset$leftright_lr4, #right left
        dataset$leftright_lr5  #right left
)
title("leftright")

# Write the PDF
dev.off()