R/kobo_crunch.R
In ElliottMess/koboloadeRlight: Data crunching light for kobo forms
#' @name kobo_crunching_all
#' @rdname kobo_crunching_all
#' @title Crunch all questions in the dictionnary
#' @description Run kobo_crunching on all the questions present in the dictionnary and the dataframe.
#'
#' @param data Dataframe containing data to analyse
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four to five named elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @export kobo_crunching_all
#'
kobo_crunching_all <- function(data, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
all_questions <- as.character(dico[dico$formpart == "questions" & dico$name %in% names(data), "name"])
all_questions_result <- all_questions %>%
map(kobo_crunching, data = data, dicofile = dicofile)
names(all_questions_result) <- all_questions
return(all_questions_result)
}
NULL
#' @name kobo_crunching
#' @rdname kobo_crunching
#' @title Pick and run analysis function for question type.
#' @description This function pick and run the right analysis function depending on the type of question for data crunching.
#'
#' @param question Question to be treated by the function. It should be the "name" of the questionas in the dictionnary or XLSform.
#' @param data Dataframe containing data to analyse
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four to five named elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @export kobo_crunching
#'
kobo_crunching <- function(question, data, dicofile = "dico.csv") {
if (is.na(question) == TRUE | is.character(question) == FALSE) stop("Please enter a question.")
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
qtype <- as.character(selectquestion$type)
if (qtype == "select_one") {
result <- kobo_bar_one_indiv(data, question, dicofile)
}else if (qtype == "select_multiple") {
result <- kobo_bar_multi_indiv(data, question, dicofile)
}else if (qtype == "integer" | qtype == "decimal" | qtype == "calculate") {
result <- kobo_histo_indiv(data, question, dicofile)
}else {
result <- list(question, "Question not treatable", NULL, NULL)
}
return(result)
}
NULL
#' @name kobo_bar_one_indiv
#' @rdname kobo_bar_one_indiv
#' @title Generate bar Chart for select_one questions
#' @description Generate basic data exploration analysis for kobo select_one question.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the question as in the dictionnary or XLSform.
#' The question has to be a select_one question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @author Edouard Legoupil, Elliott Messeiller
#' @export kobo_bar_one_indiv
#'
kobo_bar_one_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if ((selectquestion$type %in% c("select_one")) == FALSE) stop("Not a select_one question")
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
selectlistname <- as.character(selectquestion$listname)
selectchoices <- dico[dico$type == "select_one_d" & dico$listname == selectlistname, c("listname", "name", "labelchoice")]
## Matching data from question
dataquestion <- data[question]
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion)) {
stop("No answer to question")
}
## Labeling dataframe
dataquestion <- kobo_label(dataquestion, dico)
## Replacing empty cells by NAs
# dataquestion[,dataquestion==""]<-NA
### Now let's create bar graphs
variablename <- as.character(names(dataquestion))
title <- attributes(dataquestion)$variable.labels
ordinal <- as.character(selectquestion$ordinal)
if (is.na(selectfacet) == F && selectfacet != "") {
dataquestion <- cbind(dataquestion, data[selectfacet])
}
frequ <- data.frame(table(dataquestion))
if (ncol(frequ) == 2) {
names(frequ) <- c("Var1", "Freq")
} else {
names(frequ) <- c("Var1", "facet", "Freq")
frequ$facet <- as.character(frequ$facet)
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ <- frequ %>%
left_join(choices_f, by = c("facet"="name"))%>%
select(Var1, label, Freq)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet", "Freq")
}
frequ$freqper <- as.numeric(frequ$Freq / (sum(!is.na(dataquestion[1]))))
frequ$Var1 <- str_wrap(frequ$Var1, width = 15)
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
frequ$Var1 <- as.factor(frequ$Var1)
if (is.na(ordinal) == T) {
frequ$Var1 <- factor(frequ$Var1, levels = unique(frequ$Var1[order(frequ$freqper)]))
} else {
ordinal_choices <- as.character(selectchoices[, c("labelchoice")])
frequ$Var1 <- gdata::reorder.factor(frequ$Var1, new.order = ordinal_choices)
frequ %>% arrange(Var1)
}
color <- "#2a87c8"
background_rect <- data.frame(unique(frequ[, c("Var1")]))
names(background_rect) <- c("Var1")
background_rect$freqper <- 1
theme_set(theme_gray(base_size = 10))
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1, y = freqper))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_bar(data = background_rect, aes(x = Var1), stat = "identity", alpha = 0.2) +
geom_bar(stat = "identity", position = "dodge", aes(fill = facet)) +
geom_text(aes(label = paste(round(freqper * 100), "%", sep = ""), fill = facet, hjust = -0.5), position = position_dodge(width = 0.8))
} else {
plotfreq <- plotfreq +
geom_bar(fill = color, colour = color, stat = "identity") +
geom_text(aes(label = paste(round(frequ$freqper * 100), "%", sep = ""), hjust = -0.5))
}
plotfreq <- plotfreq +
scale_y_continuous(labels = percent, limits = c(0, 1.05)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
coord_flip() +
labs(x = "", y = "")+
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ
frequ_nice$freqper <- round(frequ$freqper * 100, 2)
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Choices", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Choices", "# answered", "% answered")
}
frequ_nice$Choices <- gsub(pattern = "\\n", " ", frequ_nice$Choices)
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, other = other_info)
return(out_indiv)
}
NULL
#' @name kobo_bar_multi_indiv
#' @rdname kobo_bar_multi_indiv
#' @title Generate bar Chart for select_multiple question
#' @description Generate basic data exploration analysis for kobo select_multiple question.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the questions as in the dictionnary or XLSform.
#' The question has to be a select_multiple question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @author Edouard Legoupil, Elliott Messeiller
#'
#'
kobo_bar_multi_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if (selectquestion$type != "select_multiple") stop(paste0("Not a select_multiple question. Question type in dico for this question : ", selectquestion$type))
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
selectlistname <- as.character(selectquestion$listname)
selectchoices <- dico[dico$type == "select_multiple_d" & dico$listname == selectlistname, c("listname", "name", "labelchoice")]
## Matching data from question and formatting
dataquestion <- data[, grep(question, names(data))]
dataquestion <- dataquestion %>%
select(-question) %>%
mutate_all(as.numeric)
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion) * ncol(dataquestion)) { stop("No answer to question") }
### Now let's create bar graphs
variablename <- as.character(selectquestion$name)
title <- selectquestion$label
ordinal <- as.character(selectquestion$ordinal)
# if (usedweight == "sampling_frame") {
# ### TO BE FIXED ###
# frequ <- data.frame(svytable(~dataquestion, surveydesign))
# } else {
if (is.na(selectfacet) == FALSE && selectfacet != "") {
### With a disaggregation
dataquestion <- cbind(dataquestion, data[selectfacet])
names(dataquestion)[ncol(dataquestion)] <- "facet"
labelschoices <- selectchoices %>% select(-listname)
frequ <- dataquestion %>%
gather(name, freq, -facet) %>%
dplyr::group_by(facet, name) %>%
dplyr::summarise(freq = sum(freq, na.rm = TRUE))
### Relabel CHOICES
frequ$name <- gsub(paste0(question, "."), "", frequ$name)
labelschoices$name <- as.character(labelschoices$name)
labelschoices$labelchoice <- as.character(labelschoices$labelchoice)
frequ <- frequ %>%
left_join(labelschoices, by = "name") %>%
select(labelchoice, facet, freq)
names(frequ) <- c("Var1", "facet", "Freq")
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ$facet <- as.character(frequ$facet)
frequ <- frequ %>%
left_join(choices_f, by = c("facet"="name"))%>%
ungroup()%>%
select(Var1, label, Freq)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet", "Freq")
} else {
### Without disaggregation
labelschoices <- selectchoices %>%
select(-listname) %>%
mutate_all(as.character)
frequ <- dataquestion %>%
gather(name, freq) %>%
dplyr::group_by(name) %>%
dplyr::summarise(freq = sum(freq, na.rm = TRUE))
frequ$name <- gsub(paste0(question, "."), "", frequ$name)
#### Relabel CHOICES
frequ <- frequ %>%
left_join(labelschoices, by = "name") %>%
select(labelchoice, freq)
}
#}
if (ncol(frequ) == 2) {
names(frequ) <- c("Var1", "Freq")
} else {
names(frequ) <- c("Var1", "facet", "Freq")
}
frequ$freqper <- as.numeric(frequ$Freq / (sum(!is.na(dataquestion[1]))))
frequ$Var1 <- str_wrap(frequ$Var1, width = 15)
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
frequ$Var1 <- as.factor(frequ$Var1)
if (is.na(ordinal) == T) {
frequ$Var1 <- factor(frequ$Var1, levels = unique(frequ$Var1[order(frequ$freqper)]))
} else {
ordinal_choices <- as.character(selectchoices_questions[selectchoices_questions$qname == variablename, c("labelchoice")])
frequ$Var1 <- gdata::reorder.factor(frequ$Var1, new.order = ordinal_choices)
frequ <- frequ %>% arrange(Var1)
}
theme_set(theme_gray(base_size = 15))
color <- "#2a87c8"
background_rect <- data.frame(unique(frequ[, c("Var1")]))
names(background_rect) <- c("Var1")
background_rect$freqper <- 1
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1, y = freqper))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_bar(data = background_rect, aes(x = Var1), stat = "identity", alpha = 0.2) +
geom_bar(stat = "identity", position = "dodge", aes(fill = facet)) +
geom_text(aes(label = paste(round(freqper * 100), "%", sep = ""), fill = facet, hjust = -0.5), position = position_dodge(width = 0.8))
} else {
plotfreq <- plotfreq +
geom_bar(fill = color, colour = color, stat = "identity") +
geom_text(aes(label = paste(round(frequ$freqper * 100), "%", sep = ""), hjust = -0.5))
}
plotfreq <- plotfreq +
scale_y_continuous(labels = percent, limits = c(0, 1.05)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
coord_flip() +
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ
frequ_nice$freqper <- round(frequ$freqper * 100, 2)
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Choices", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Choices", "# answered", "% answered")
}
frequ_nice$Choices <- gsub(pattern = "\\n", " ", frequ_nice$Choices)
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, other = other_info)
return(out_indiv)
}
NULL
#' @name kobo_histo_indiv
#' @rdname kobo_histo_indiv
#' @title Generate histograme for individual integer questions
#'
#' @description Automatically generate histogrammes for each of the integer questions in the dataset. ggplot2 is used.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the questions as in the dictionnary or XLSform.
#' The question has to be a decimal or integer question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#'
#' @author Edouard Legoupil, Elliott Messeiller
#'
#' @examples
#' kobo_histo_indiv()
#' @export kobo_histo_indiv
#'
#' @examples
#' \dontrun{
#' kobo_histo_indiv()
#' }
#'
kobo_histo_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if ((selectquestion$type %in% c("integer", "decimal", "calculate")) == FALSE) stop("Not a question")
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
## Matching data from question
dataquestion <- data[question]
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion)) { stop("No answer to question") }
## Labeling dataframe
dataquestion <- kobo_label(dataquestion, dico)
## Replacing empty cells by NAs
# dataquestion[,dataquestion==""]<-NA
### Now let's create bar graphs
variablename <- as.character(names(dataquestion))
title <- attributes(dataquestion)$variable.labels
ordinal <- as.character(dico[dico$name == variablename, c("ordinal")])
if (is.na(selectfacet) == F && selectfacet != "") {
dataquestion <- cbind(dataquestion, data[selectfacet])
}
#if (usedweight == "sampling_frame") {
### TO BE FIXED ###
# frequ <- data.frame(svytable(~dataquestion, surveydesign))
# } else {
# frequ <- dataquestion
# }
frequ <- dataquestion
if (ncol(frequ) == 1) {
names(frequ) <- "Var1"
frequ$Var1 <- as.numeric(as.character(frequ$Var1))
} else {
names(frequ) <- c("Var1", "name")
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ$name <- as.character(frequ$name)
frequ <- frequ %>%
left_join(choices_f, by = "name")%>%
select(Var1, label)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet")
}
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
color <- "#2a87c8"
theme_set(theme_gray(base_size = 10))
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_density(aes(fill = facet), adjust = 2, alpha = 0.2)
} else {
plotfreq <- plotfreq +
geom_density(adjust = 2, alpha = 0.2)
}
plotfreq <- plotfreq +
scale_x_continuous(expand = c(0, 0)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ %>%
table() %>%
data.frame() %>%
mutate(freqper = paste0(round(Freq / sum(Freq) * 100), "%"))
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Variable", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Variable", "# answered", "% answered")
}
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, summary = summary(frequ), other = other_info)
return(out_indiv)
}
NULL
ElliottMess/koboloadeRlight documentation built on June 20, 2019, 2:24 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#' @name kobo_crunching_all
#' @rdname kobo_crunching_all
#' @title Crunch all questions in the dictionnary
#' @description Run kobo_crunching on all the questions present in the dictionnary and the dataframe.
#'
#' @param data Dataframe containing data to analyse
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four to five named elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @export kobo_crunching_all
#'
kobo_crunching_all <- function(data, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
all_questions <- as.character(dico[dico$formpart == "questions" & dico$name %in% names(data), "name"])
all_questions_result <- all_questions %>%
map(kobo_crunching, data = data, dicofile = dicofile)
names(all_questions_result) <- all_questions
return(all_questions_result)
}
NULL
#' @name kobo_crunching
#' @rdname kobo_crunching
#' @title Pick and run analysis function for question type.
#' @description This function pick and run the right analysis function depending on the type of question for data crunching.
#'
#' @param question Question to be treated by the function. It should be the "name" of the questionas in the dictionnary or XLSform.
#' @param data Dataframe containing data to analyse
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four to five named elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @export kobo_crunching
#'
kobo_crunching <- function(question, data, dicofile = "dico.csv") {
if (is.na(question) == TRUE | is.character(question) == FALSE) stop("Please enter a question.")
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
qtype <- as.character(selectquestion$type)
if (qtype == "select_one") {
result <- kobo_bar_one_indiv(data, question, dicofile)
}else if (qtype == "select_multiple") {
result <- kobo_bar_multi_indiv(data, question, dicofile)
}else if (qtype == "integer" | qtype == "decimal" | qtype == "calculate") {
result <- kobo_histo_indiv(data, question, dicofile)
}else {
result <- list(question, "Question not treatable", NULL, NULL)
}
return(result)
}
NULL
#' @name kobo_bar_one_indiv
#' @rdname kobo_bar_one_indiv
#' @title Generate bar Chart for select_one questions
#' @description Generate basic data exploration analysis for kobo select_one question.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the question as in the dictionnary or XLSform.
#' The question has to be a select_one question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @author Edouard Legoupil, Elliott Messeiller
#' @export kobo_bar_one_indiv
#'
kobo_bar_one_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if ((selectquestion$type %in% c("select_one")) == FALSE) stop("Not a select_one question")
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
selectlistname <- as.character(selectquestion$listname)
selectchoices <- dico[dico$type == "select_one_d" & dico$listname == selectlistname, c("listname", "name", "labelchoice")]
## Matching data from question
dataquestion <- data[question]
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion)) {
stop("No answer to question")
}
## Labeling dataframe
dataquestion <- kobo_label(dataquestion, dico)
## Replacing empty cells by NAs
# dataquestion[,dataquestion==""]<-NA
### Now let's create bar graphs
variablename <- as.character(names(dataquestion))
title <- attributes(dataquestion)$variable.labels
ordinal <- as.character(selectquestion$ordinal)
if (is.na(selectfacet) == F && selectfacet != "") {
dataquestion <- cbind(dataquestion, data[selectfacet])
}
frequ <- data.frame(table(dataquestion))
if (ncol(frequ) == 2) {
names(frequ) <- c("Var1", "Freq")
} else {
names(frequ) <- c("Var1", "facet", "Freq")
frequ$facet <- as.character(frequ$facet)
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ <- frequ %>%
left_join(choices_f, by = c("facet"="name"))%>%
select(Var1, label, Freq)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet", "Freq")
}
frequ$freqper <- as.numeric(frequ$Freq / (sum(!is.na(dataquestion[1]))))
frequ$Var1 <- str_wrap(frequ$Var1, width = 15)
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
frequ$Var1 <- as.factor(frequ$Var1)
if (is.na(ordinal) == T) {
frequ$Var1 <- factor(frequ$Var1, levels = unique(frequ$Var1[order(frequ$freqper)]))
} else {
ordinal_choices <- as.character(selectchoices[, c("labelchoice")])
frequ$Var1 <- gdata::reorder.factor(frequ$Var1, new.order = ordinal_choices)
frequ %>% arrange(Var1)
}
color <- "#2a87c8"
background_rect <- data.frame(unique(frequ[, c("Var1")]))
names(background_rect) <- c("Var1")
background_rect$freqper <- 1
theme_set(theme_gray(base_size = 10))
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1, y = freqper))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_bar(data = background_rect, aes(x = Var1), stat = "identity", alpha = 0.2) +
geom_bar(stat = "identity", position = "dodge", aes(fill = facet)) +
geom_text(aes(label = paste(round(freqper * 100), "%", sep = ""), fill = facet, hjust = -0.5), position = position_dodge(width = 0.8))
} else {
plotfreq <- plotfreq +
geom_bar(fill = color, colour = color, stat = "identity") +
geom_text(aes(label = paste(round(frequ$freqper * 100), "%", sep = ""), hjust = -0.5))
}
plotfreq <- plotfreq +
scale_y_continuous(labels = percent, limits = c(0, 1.05)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
coord_flip() +
labs(x = "", y = "")+
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ
frequ_nice$freqper <- round(frequ$freqper * 100, 2)
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Choices", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Choices", "# answered", "% answered")
}
frequ_nice$Choices <- gsub(pattern = "\\n", " ", frequ_nice$Choices)
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, other = other_info)
return(out_indiv)
}
NULL
#' @name kobo_bar_multi_indiv
#' @rdname kobo_bar_multi_indiv
#' @title Generate bar Chart for select_multiple question
#' @description Generate basic data exploration analysis for kobo select_multiple question.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the questions as in the dictionnary or XLSform.
#' The question has to be a select_multiple question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#' @author Edouard Legoupil, Elliott Messeiller
#'
#'
kobo_bar_multi_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
# Select question in dico
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if (selectquestion$type != "select_multiple") stop(paste0("Not a select_multiple question. Question type in dico for this question : ", selectquestion$type))
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
selectlistname <- as.character(selectquestion$listname)
selectchoices <- dico[dico$type == "select_multiple_d" & dico$listname == selectlistname, c("listname", "name", "labelchoice")]
## Matching data from question and formatting
dataquestion <- data[, grep(question, names(data))]
dataquestion <- dataquestion %>%
select(-question) %>%
mutate_all(as.numeric)
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion) * ncol(dataquestion)) { stop("No answer to question") }
### Now let's create bar graphs
variablename <- as.character(selectquestion$name)
title <- selectquestion$label
ordinal <- as.character(selectquestion$ordinal)
# if (usedweight == "sampling_frame") {
# ### TO BE FIXED ###
# frequ <- data.frame(svytable(~dataquestion, surveydesign))
# } else {
if (is.na(selectfacet) == FALSE && selectfacet != "") {
### With a disaggregation
dataquestion <- cbind(dataquestion, data[selectfacet])
names(dataquestion)[ncol(dataquestion)] <- "facet"
labelschoices <- selectchoices %>% select(-listname)
frequ <- dataquestion %>%
gather(name, freq, -facet) %>%
dplyr::group_by(facet, name) %>%
dplyr::summarise(freq = sum(freq, na.rm = TRUE))
### Relabel CHOICES
frequ$name <- gsub(paste0(question, "."), "", frequ$name)
labelschoices$name <- as.character(labelschoices$name)
labelschoices$labelchoice <- as.character(labelschoices$labelchoice)
frequ <- frequ %>%
left_join(labelschoices, by = "name") %>%
select(labelchoice, facet, freq)
names(frequ) <- c("Var1", "facet", "Freq")
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ$facet <- as.character(frequ$facet)
frequ <- frequ %>%
left_join(choices_f, by = c("facet"="name"))%>%
ungroup()%>%
select(Var1, label, Freq)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet", "Freq")
} else {
### Without disaggregation
labelschoices <- selectchoices %>%
select(-listname) %>%
mutate_all(as.character)
frequ <- dataquestion %>%
gather(name, freq) %>%
dplyr::group_by(name) %>%
dplyr::summarise(freq = sum(freq, na.rm = TRUE))
frequ$name <- gsub(paste0(question, "."), "", frequ$name)
#### Relabel CHOICES
frequ <- frequ %>%
left_join(labelschoices, by = "name") %>%
select(labelchoice, freq)
}
#}
if (ncol(frequ) == 2) {
names(frequ) <- c("Var1", "Freq")
} else {
names(frequ) <- c("Var1", "facet", "Freq")
}
frequ$freqper <- as.numeric(frequ$Freq / (sum(!is.na(dataquestion[1]))))
frequ$Var1 <- str_wrap(frequ$Var1, width = 15)
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
frequ$Var1 <- as.factor(frequ$Var1)
if (is.na(ordinal) == T) {
frequ$Var1 <- factor(frequ$Var1, levels = unique(frequ$Var1[order(frequ$freqper)]))
} else {
ordinal_choices <- as.character(selectchoices_questions[selectchoices_questions$qname == variablename, c("labelchoice")])
frequ$Var1 <- gdata::reorder.factor(frequ$Var1, new.order = ordinal_choices)
frequ <- frequ %>% arrange(Var1)
}
theme_set(theme_gray(base_size = 15))
color <- "#2a87c8"
background_rect <- data.frame(unique(frequ[, c("Var1")]))
names(background_rect) <- c("Var1")
background_rect$freqper <- 1
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1, y = freqper))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_bar(data = background_rect, aes(x = Var1), stat = "identity", alpha = 0.2) +
geom_bar(stat = "identity", position = "dodge", aes(fill = facet)) +
geom_text(aes(label = paste(round(freqper * 100), "%", sep = ""), fill = facet, hjust = -0.5), position = position_dodge(width = 0.8))
} else {
plotfreq <- plotfreq +
geom_bar(fill = color, colour = color, stat = "identity") +
geom_text(aes(label = paste(round(frequ$freqper * 100), "%", sep = ""), hjust = -0.5))
}
plotfreq <- plotfreq +
scale_y_continuous(labels = percent, limits = c(0, 1.05)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
coord_flip() +
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ
frequ_nice$freqper <- round(frequ$freqper * 100, 2)
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Choices", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Choices", "# answered", "% answered")
}
frequ_nice$Choices <- gsub(pattern = "\\n", " ", frequ_nice$Choices)
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, other = other_info)
return(out_indiv)
}
NULL
#' @name kobo_histo_indiv
#' @rdname kobo_histo_indiv
#' @title Generate histograme for individual integer questions
#'
#' @description Automatically generate histogrammes for each of the integer questions in the dataset. ggplot2 is used.
#'
#' @param data Dataframe containing data to analyse
#' @param question Question to be treated by the function. It should be the "name" of the questions as in the dictionnary or XLSform.
#' The question has to be a decimal or integer question in the dictionary.
#' @param dicofile Name of the file containing the dictionnary. "dico.csv" is used by default.
#'
#' @return Returns a list containing four elements:
#' \itemize{
#' \item variable : The name of the question (variable) analysed
#' \item plot: A ggplot2 object with a basic bar or density graph depending on the question
#' \item freq_tbl: A frequency table of the question of the question
#' \item summary: summary statistics (quartiles and mean). Only for decimal/integer questions.
#' \item other: Other information on the question: # respondents answered the form, # and % of respondents to the question.
#'}
#'
#'
#' @author Edouard Legoupil, Elliott Messeiller
#'
#' @examples
#' kobo_histo_indiv()
#' @export kobo_histo_indiv
#'
#' @examples
#' \dontrun{
#' kobo_histo_indiv()
#' }
#'
kobo_histo_indiv <- function(data, question, dicofile = "dico.csv") {
dico <- read.csv(dicofile)
selectquestion <- filter(dico, name == question)
# Checks
## Check that variable are in the dataset
if (nrow(selectquestion) != 1) stop("Question not in dictionnary", call. = FALSE)
## Check that it is a select_one question
if ((selectquestion$type %in% c("integer", "decimal", "calculate")) == FALSE) stop("Not a question")
## Select disaggregation if exists
if(selectquestion$disaggregation == "" || is.na(selectquestion$disaggregation)){
selectfacet <- NA
}else{
selectfacet <- as.character(selectquestion$disaggregation)
}
if(selectfacet %in% names(data) == FALSE & is.na(selectfacet) == FALSE) stop ("Disaggregation not in columns. Please enter a column as disaggregation")
## Getting choices
## Matching data from question
dataquestion <- data[question]
## Checking that question was answered
if (sum(is.na(dataquestion)) == nrow(dataquestion)) { stop("No answer to question") }
## Labeling dataframe
dataquestion <- kobo_label(dataquestion, dico)
## Replacing empty cells by NAs
# dataquestion[,dataquestion==""]<-NA
### Now let's create bar graphs
variablename <- as.character(names(dataquestion))
title <- attributes(dataquestion)$variable.labels
ordinal <- as.character(dico[dico$name == variablename, c("ordinal")])
if (is.na(selectfacet) == F && selectfacet != "") {
dataquestion <- cbind(dataquestion, data[selectfacet])
}
#if (usedweight == "sampling_frame") {
### TO BE FIXED ###
# frequ <- data.frame(svytable(~dataquestion, surveydesign))
# } else {
# frequ <- dataquestion
# }
frequ <- dataquestion
if (ncol(frequ) == 1) {
names(frequ) <- "Var1"
frequ$Var1 <- as.numeric(as.character(frequ$Var1))
} else {
names(frequ) <- c("Var1", "name")
listname_f <- as.character(dico[dico$name == selectfacet & dico$formpart == "questions", "listname" ])
choices_f <- unique(dico[dico$listname == listname_f & dico$formpart == "answers", c("name", "label")])
choices_f$name <- as.character(choices_f$name)
frequ$name <- as.character(frequ$name)
frequ <- frequ %>%
left_join(choices_f, by = "name")%>%
select(Var1, label)
frequ <- droplevels(frequ)
names(frequ) <- c("Var1", "facet")
}
totalanswer <- nrow(dataquestion)
count_replied <- (sum(!is.na(dataquestion[1])))
percentresponse <- paste(round((count_replied / totalanswer * 100), digits = 2), "%", sep = "")
color <- "#2a87c8"
theme_set(theme_gray(base_size = 10))
## and now the graph
plotfreq <- ggplot(frequ, aes(x = Var1))
if ("facet" %in% colnames(frequ)) {
plotfreq <- plotfreq +
geom_density(aes(fill = facet), adjust = 2, alpha = 0.2)
} else {
plotfreq <- plotfreq +
geom_density(adjust = 2, alpha = 0.2)
}
plotfreq <- plotfreq +
scale_x_continuous(expand = c(0, 0)) +
scale_fill_brewer(palette = "PuBu", name = "Disaggregation") +
labs(x = "", y = "")+
ggtitle(str_wrap(title, width = 50)) +
theme(
plot.title = element_text(face = "bold", size = 20),
plot.background = element_rect(fill = "transparent", colour = NA)
)
## Formating the frequ table
frequ_nice <- frequ %>%
table() %>%
data.frame() %>%
mutate(freqper = paste0(round(Freq / sum(Freq) * 100), "%"))
if(ncol(frequ_nice) == 4){
names(frequ_nice) <- c("Variable", "Disaggregation", "# answered", "% answered")
}else{
names(frequ_nice) <- c("Variable", "# answered", "% answered")
}
other_info <- data.frame(
title = c("# Answered form", "# Answered question", "% Answered question"),
info = c(totalanswer, count_replied, percentresponse)
)
## Saving outputs to a list
out_indiv <- list(variable = variablename, plot = plotfreq, freq_tbl = frequ_nice, summary = summary(frequ), other = other_info)
return(out_indiv)
}
NULL
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.