R/qog.R
In briatte/qogdata: Functions to manipulate Quality of Government data and related material
#' Get Quality of Government datasets in \code{xtdata} format
#'
#' Function to download Quality of Government (QOG) data and load it as a data
#' frame in R. The result carries an \code{\link{xtdata}} attribute that can be
#' passed to the \code{\link{xtmerge}} panel data method. Please visit the
#' QOG Institute website at \url{http://www.qog.pol.gu.se/} for a presentation
#' of QOG research.
#'
#' @export
#' @aliases get_qog
#' @param file a filename to save the dataset at.
#' If set to \code{FALSE} (the default), the function just returns the link
#' to the dataset.
#' If set to \code{TRUE}, the server filename of the dataset is used, which
#' returns either a CSV file if \code{version} is set to \code{std}, or
#' a Stata \code{dta} file otherwise. See 'Details'.
#' @param replace whether to overwrite the dataset even if a file already
#' exists at the download location. Defaults to \code{FALSE}.
#' @param path a folder path to prepend to the filename and to the codebook
#' if \code{codebook} is not \code{FALSE}.
#' @param version the QOG version:
#' \code{std} (Standard), \code{soc} (Social Policy), \code{bas} (Basic)
#' or \code{exp} (Expert). Defaults to \code{std}. See 'Details'.
#' @param format the QOG format, usually \code{cs} for cross-sectional data
#' or \code{ts} for time series in the \code{std} and \code{bas} versions.
#' See 'Details' for the full list of specifications. Defaults to \code{cs}.
#' @param codebook whether to download the codebook. Calls \code{qogbook} by
#' passing the \code{codebook}, \code{version} and \code{path} arguments to it,
#' where \code{codebook} is treated as the filename for the codebook.
#' Defaults to \code{FALSE}.
#' @param variables a selection of variables to import. \code{ccode} ISO-3N
#' country codes \code{ccode} and \code{year} identifiers will be forced into
#' the output if relevant.
#' @param years a selection of years to import. Effective only with
#' the \code{ts}, \code{tsl} or \code{ind} formats.
#' @param ... other arguments supplied to the import method, which is
#' \code{read.csv} by default,
#' or \code{\link[foreign]{read.dta}} if \code{file} is a Stata \code{dta} dataset,
#' or \code{\link[foreign]{read.spss}} if \code{file} is a SPSS \code{sav} dataset.
#' @details This version of the package handles all four QOG datasets:
#' \tabular{lcl}{
#' QOG Standard \tab \code{std} \tab 15 May 2013\cr
#' QOG Social Policy \tab \code{soc} \tab 4 April 2012\cr
#' QOG Basic \tab \code{bas}): \tab 28 March 2011\cr
#' QOG Expert Survey \tab \code{exp} \tab 3-6 September 2012\cr
#' URL: \tab \tab \url{http://www.qog.pol.gu.se}\cr
#' }
#'
#' Each QOG dataset is available in a variety of data formats:
#'
#' \itemize{
#' \item QOG datasets \code{std} and \code{bas}
#' require format \code{cs} (cross-section)
#' or \code{ts} (time series).
#' \item QOG dataset \code{soc}
#' requires format \code{cs}, \code{tsl} (time series, long)
#' or \code{tsw} (time series, wide)
#' \item QOG dataset \code{exp}
#' requires format \code{cntry} (country-level)
#' or \code{ind} (individual survey)
#' }
#'
#' The QOG Standard series comes in CSV, SPSS and Stata file formats, CVS being
#' the only format that contains numeric codes instead of QOG value labels.
#' Datasets outside of the QOG Standard series are available only as Stata items
#' and require that \code{file} ends in \code{.dta} when \code{version} is not
#' \code{std}. The only exception is dataset \code{csyom}, which automatically
#' sets \code{version} to \code{std} and requires \code{file} to end
#' in \code{.csv}. Filenames with inadequate extensions will be modified to
#' conform to these expectations if they do not.
#' @seealso \code{\link{qogbook}}, \code{\link[foreign]{read.dta}}, \code{\link[foreign]{read.spss}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # Show URL to QOG Standard cross-section.
#' qogdata()
#' # Show URL to QOG Social Policy time series, long format.
#' qogdata(version = "soc", format = "tsl")
#' ## Download codebook and recent years from QOG Basic cross-section (not run).
#' # QOG = qogdata(file = "qog.cs.txt", version = "bas", format = "cs",
#' # years = 2002:2012, codebook = TRUE)
#' ## Download QOG Standard cross-section years of measurement (not run).
#' # QOG = qogdata(tempfile(fileext = ".csv"), format = "csyom")
#' ## Show QOG years of measurement for Gini coefficient (not run).
#' # table(QOG$wdi_gini)
#' @keywords qog
qogdata <- function(file = FALSE, replace = FALSE, codebook = FALSE, path = "",
version = "std", format = "cs",
variables = NULL, years = NULL, ...) {
try_require("foreign")
#
# currently available
#
versions = list(
std = c("ts", "cs", "csyom"),
bas = c("ts", "cs"),
exp = c("ctry", "ind"),
soc = c("cs", "tsl", "tsw"))
#
# correct version
#
if(!version %in% names(versions)) {
stop("Invalid version: use one of ",
paste0(names(versions), collapse = ", "))
}
#
# correct format
#
if(format == "csyom") {
version = "std"
if(!grepl(".csv$|.txt$", file))
file = gsub("(\\.|\\w){4}$", ".csv", file)
}
if(!format %in% unlist(versions[version])) {
stop("Invalid format: use one of ",
paste0(unlist(versions[version]), collapse = ", "))
}
#
# automatic filename
#
if(isTRUE(file)) {
file = paste0("qog_",
version,
"_",
format,
ifelse(version == "std",
paste0("_", "15May13.csv"),
".dta")
)
}
else {
if(is.character(file) & version != "std" & !grepl(".dta$", file)) {
file = gsub("(\\.|\\w){4}$", ".dta", file)
warning("QOG datasets other than std are available only as Stata files.\n",
" The filename that you specified was modified to ", file)
}
}
if(is.character(path))
if(nchar(path) > 0) file = paste(path, file, sep = "/")
#
# online source
#
link = paste0("http://www.qogdata.pol.gu.se/data/",
ifelse(version == "std", "QoG", "qog"),
"_", version, "_", format,
ifelse(version == "std", paste0("_", "15May13"), ""),
ifelse(version == "std" & grepl("csv|dta|sav", file),
substring(file, nchar(file) - 3),
".dta")
)
if(is.logical(file)) {
return(link)
}
else {
if(replace || !file.exists(file)) {
message("Downloading ", link, "...")
download.file(link, file, mode = "wb", quiet = TRUE)
}
else {
message("Loading from disk...")
}
}
#
# reader call
#
read = "read.csv"
args = list(file = file, ...)
# foreign or not
if(!grepl(".dta$|.sav$", file))
args["sep"] = ";"
# stata args
if(grepl(".dta$", file)) {
read = "read.dta"
if(is.null(unlist(args["warn.missing.labels"])))
args["warn.missing.labels"] = FALSE
}
# spss args
if(grepl(".sav$", file)) {
read = "read.spss"
if(is.null(unlist(args["to.data.frame"])))
args["to.data.frame"] = TRUE
}
data = do.call(read, args)
#
# selected variables
#
uids = c("ccode", "ccodealp", "cname", "ccodecow", "ccodewb")
pids = uids %in% names(data)
# avoid ts bug
if(grepl("ts|tsl", format) & !any(pids))
stop("You are trying to load a QOG dataset as time series, but it has no identifier variable.")
# avoid ts bug
if(grepl("ts|tsl", format) & !"year" %in% names(data))
stop("You are trying to load a QOG dataset as time series, but it has no year variable.")
if(!is.null(variables)) {
if(grepl("ts|tsl", format) & !"year" %in% variables) {
warning("Forcing year identifier into the dataset.")
variables = c("year", variables)
}
if(grepl("std|bas|soc", version) & !"ccode" %in% variables) {
warning("Forcing ccode identifier into the dataset.")
variables = c("ccode", variables)
}
data = data[, names(data) %in% variables]
}
#
# selected years
#
if(!is.null(years) && format %in% c("ts", "tsl", "ind"))
data = data[data$year %in% years, ]
#
# message
#
message("Loaded ", file, " (N = ", nrow(data),
ifelse(format %in% c("ts", "tsl", "ind"),
paste0(", ", min(data$year),
"-", max(data$year),
", T = ", length(unique(data$year))),
""),
").")
#
# grab codebook
#
if(isTRUE(codebook) || grepl(".pdf", codebook))
qogbook(codebook, version, path, replace)
#
# xtdata spec
#
pids = uids[pids]
if(format == "ts" | format == "tsl") {
data = xtset(data,
data = c(pids[1], "year", pids[-1]),
spec = c("iso3n", "year"),
type = "country",
name = "Quality of Government, time series data"
)
}
#
# finish line
#
return(data)
}
#' Download Quality of Government codebooks
#'
#' Function to download Quality of Government (QOG) codebooks. Please visit
#' the QOG Institute website at \url{http://www.qog.pol.gu.se/} for a
#' presentation of QOG research.
#'
#' @export
#' @param file a filename to save the codebook at.
#' If set to \code{TRUE}, the name of the codebook on the QOG server will be used.
#' If set to \code{FALSE} (the default), the function only returns the link to
#' the dataset. The filename must end in \code{.pdf}.
#' @param replace whether to download the dataset even if a file already exists
#' at the download location. Defaults to \code{FALSE}.
#' @param path a folder path to append to the filename.
#' @param version the QOG version: \code{std} (Standard),
#' \code{soc} (Social Policy), \code{bas} (Basic) or \code{exp} (Expert).
#' Defaults to \code{std}.
#' @details The function mimics Richard Svensson's \code{qogbook} Stata command.
#' @seealso \code{\link{qogdata}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # Show the URL to the QOG Standard dataset codebook.
#' qogbook()
#' ## Download QOG Standard codebook with default filename (not run).
#' # qogbook(file = TRUE)
#' ## Download QOG Basic dataset codebook to specific filename (not run).
#' # qogbook(file = "qog.basic.codebook.pdf", version = "bas")
#' @keywords qog
qogbook <- function(file = FALSE, version = "std", path = "", replace = FALSE) {
if(!version %in% c("std", "bas"))
stop("Codebook available only for versions bas, std")
link = "http://www.qogdata.pol.gu.se/data/Codebook_QoG_Std15May13.pdf"
if(version == "bas") {
link = "http://www.qogdata.pol.gu.se/codebook/codebook_basic_20120608.pdf"
}
if(isTRUE(file) & version == "bas") {
file = "codebook_basic_20120608.pdf"
}
else if(isTRUE(file)) {
file = "Codebook_QoG_Std15May13.pdf"
}
#
# path
#
if(is.character(path))
if(nchar(path) > 0) file = paste(path, file, sep = "/")
#
# download
#
if(is.logical(file)) {
return(link)
}
else if(!grepl(".pdf$", file)) {
stop("Please specify a .pdf codebook filename or set file = TRUE.")
}
else {
if(replace || !file.exists(file)) {
message("Downloading codebook to ", file, "...")
download.file(link, file, mode = "wb", quiet = TRUE)
}
message("Codebook: ", file)
}
}
#' Find Quality of Government variables
#'
#' Function to perform a \code{regex} search on QOG variable names and labels.
#' A few labels are missing for strictly cross-sectional variables.
#'
#' @export
#' @param ... keywords or \code{regex} phrases passed to \code{grepl}.
#' @param version the QOG version to search: either \code{std} (the default) or \code{soc}.
#' @param compact whether to limit the labels returned to 32 characters. Defaults to \code{TRUE} for better console output.
#' @param show which variables to show for years of measurement: \code{cs} (cross-sectional), \code{ts} (time series), or \code{all} (the default).
#' @return a data frame containg matching variables, described by their names, labels and years of measurement in the time series (\code{ts}) cross-sectional (\code{cs}) datasets. The information should match the ranges indicated in the \emph{QOG Standard Codebook} and \emph{QOG Social Policy Codebook}.
#' @references
#' Svensson, Richard, Stefan Dahlberg, Staffan Kumlin & Bo Rothstein.
#' 2012. \emph{The QoG Social Policy Dataset}, version 4Apr12.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#'
#' Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Soren Holmberg,
#' Bo Rothstein, Petrus Sundin & Richard Svensson. 2013.
#' \emph{The Quality of Government Dataset}, version 15May13.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#' @seealso \code{\link{grep}}, \code{\link{qogdata}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # QOG Standard search.
#' qogfind("regime", "institutions")
#' # QOG Standard search, with regex syntax.
#' qogfind("public|administration")
#' # QOG Social Policy search, showing cross-sectional information only.
#' head(qogfind("^socx", version = "soc", show = "cs", compact = FALSE))
#' # QOG Standard variables featured only in the cross-sectional version.
#' qogfind("*")[is.na(qogfind("*")$ts.N), ]
#' @keywords qog
qogfind <- function(..., version = "std", compact = TRUE, show = "all") {
x = paste0(c(...), collapse = "|")
data(qog.index)
if (version == "std") {
message("QOG Standard results")
r = qog.std.index[grepl(x, qog.std.index$variable, ignore.case = TRUE) |
grepl(x, qog.std.index$label, ignore.case = TRUE), ]
}
if (version == "soc") {
message("QOG Social Policy results")
r = qog.soc.index[grepl(x, qog.soc.index$variable, ignore.case = TRUE) |
grepl(x, qog.soc.index$label, ignore.case = TRUE), ]
}
if (version == "bas") {
message("QOG Basic results")
r = qog.bas.index[grepl(x, qog.bas.index$variable, ignore.case = TRUE) |
grepl(x, qog.bas.index$label, ignore.case = TRUE), ]
}
r$variable = as.character(r$variable)
if(compact) r$label = substr(r$label, 1, 32)
if(show == "cs") r = r[, !grepl("ts.", names(r))]
if(show == "ts") r = r[, !grepl("cs.", names(r))]
return(r)
}
#' Join historical and recent states in QOG Standard time series data
#'
#' Function to plot maps of Quality of Government (QOG) data. Requires the \code{ggplot2} and \code{maps} packages.
#'
#' @export
#' @param data a QOG Standard time series dataset, or any data frame with \code{cname} (country) and \code{year} information coded as in the QOG Standard time series dataset.
#' @param country the country name to join data over. Requires the \code{cname} variable.
#' @details The function will try to find two series of country-year observations that both match the \code{country} argument. Within the QOG Standard time series dataset, this will match historical states like "France (-1962)" to modern states like "France (1963-)". The function will then create a new variable out of both series, joined at their separation years, and set its country code attributes to the most recent ones. See Appendix A of the \emph{QOG Standard Codebook} for details on historical states in the QOG Standard time series dataset.
#' @return a data frame with country-year observations
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @references Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Soren Holmberg,
#' Bo Rothstein, Petrus Sundin & Richard Svensson. 2013.
#' \emph{The Quality of Government Dataset}, version 15May13.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#' @examples
#' # Load QOG demo datasets.
#' data(qog.demo)
#' QOG = qog.ts.demo
#' QOG = qogjoin(QOG, "Ethiopia")
#' QOG = qogjoin(QOG, "France")
#' QOG = qogjoin(QOG, "Malaysia")
#' QOG = qogjoin(QOG, "Pakistan")
#' @keywords qog
qogjoin <- function(data, country = NULL) {
stopifnot("cname" %in% names(data))
stopifnot("year" %in% names(data))
stopifnot(is.character(country))
x = data$cname[grepl(country, data$cname)]
x = unique(x)
if(length(x) == 0)
stop("No country match for ", country)
else if(length(x) == 1)
stop("Single country match for ", country)
else if(length(x) > 2)
stop("More than two country matches: ", paste0(x, collapse = ", "))
t = xt(data)
y = as.numeric(gsub("\\D", "", x))
min = x[order(y) == 1]
max = x[order(y) == 2]
message("Joining ", min, " to ", max)
one = data[data$cname == min & data$year <= y[order(y) == 1], ]
two = data[data$cname == max & data$year >= y[order(y) == 2], ]
new = rbind(one, two)
new$cname = country
# country codes
if("ccode" %in% names(data))
new$ccode = unique(data$ccode[data$cname == max])
if("ccodealp" %in% names(data))
new$ccodealp = unique(data$ccodealp[data$cname == max])
if("cname_year" %in% names(data))
new$cname_year = paste(new$cname, new$year)
if("ccodealp_year" %in% names(data))
new$ccodealp_year = paste(new$ccodealp, new$year)
if("ccodewb" %in% names(data))
new$ccodewb = unique(data$ccodewb[data$cname == max])
message(country, " now runs over ", nrow(new), " country-year observations, ",
paste0(range(new$year), collapse = "-"))
data[, "cname"] = as.character(data[, "cname"])
data[data$cname == max, ] = new
data = subset(data, cname != min)
data[, "cname"] = factor(data[, "cname"])
data = xtset(data,
data = t$data,
spec = t$spec,
name = t$name,
url = t$url,
quiet = FALSE)
return(data)
}
briatte/qogdata documentation built on May 13, 2019, 7:43 a.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.
#' Get Quality of Government datasets in \code{xtdata} format
#'
#' Function to download Quality of Government (QOG) data and load it as a data
#' frame in R. The result carries an \code{\link{xtdata}} attribute that can be
#' passed to the \code{\link{xtmerge}} panel data method. Please visit the
#' QOG Institute website at \url{http://www.qog.pol.gu.se/} for a presentation
#' of QOG research.
#'
#' @export
#' @aliases get_qog
#' @param file a filename to save the dataset at.
#' If set to \code{FALSE} (the default), the function just returns the link
#' to the dataset.
#' If set to \code{TRUE}, the server filename of the dataset is used, which
#' returns either a CSV file if \code{version} is set to \code{std}, or
#' a Stata \code{dta} file otherwise. See 'Details'.
#' @param replace whether to overwrite the dataset even if a file already
#' exists at the download location. Defaults to \code{FALSE}.
#' @param path a folder path to prepend to the filename and to the codebook
#' if \code{codebook} is not \code{FALSE}.
#' @param version the QOG version:
#' \code{std} (Standard), \code{soc} (Social Policy), \code{bas} (Basic)
#' or \code{exp} (Expert). Defaults to \code{std}. See 'Details'.
#' @param format the QOG format, usually \code{cs} for cross-sectional data
#' or \code{ts} for time series in the \code{std} and \code{bas} versions.
#' See 'Details' for the full list of specifications. Defaults to \code{cs}.
#' @param codebook whether to download the codebook. Calls \code{qogbook} by
#' passing the \code{codebook}, \code{version} and \code{path} arguments to it,
#' where \code{codebook} is treated as the filename for the codebook.
#' Defaults to \code{FALSE}.
#' @param variables a selection of variables to import. \code{ccode} ISO-3N
#' country codes \code{ccode} and \code{year} identifiers will be forced into
#' the output if relevant.
#' @param years a selection of years to import. Effective only with
#' the \code{ts}, \code{tsl} or \code{ind} formats.
#' @param ... other arguments supplied to the import method, which is
#' \code{read.csv} by default,
#' or \code{\link[foreign]{read.dta}} if \code{file} is a Stata \code{dta} dataset,
#' or \code{\link[foreign]{read.spss}} if \code{file} is a SPSS \code{sav} dataset.
#' @details This version of the package handles all four QOG datasets:
#' \tabular{lcl}{
#' QOG Standard \tab \code{std} \tab 15 May 2013\cr
#' QOG Social Policy \tab \code{soc} \tab 4 April 2012\cr
#' QOG Basic \tab \code{bas}): \tab 28 March 2011\cr
#' QOG Expert Survey \tab \code{exp} \tab 3-6 September 2012\cr
#' URL: \tab \tab \url{http://www.qog.pol.gu.se}\cr
#' }
#'
#' Each QOG dataset is available in a variety of data formats:
#'
#' \itemize{
#' \item QOG datasets \code{std} and \code{bas}
#' require format \code{cs} (cross-section)
#' or \code{ts} (time series).
#' \item QOG dataset \code{soc}
#' requires format \code{cs}, \code{tsl} (time series, long)
#' or \code{tsw} (time series, wide)
#' \item QOG dataset \code{exp}
#' requires format \code{cntry} (country-level)
#' or \code{ind} (individual survey)
#' }
#'
#' The QOG Standard series comes in CSV, SPSS and Stata file formats, CVS being
#' the only format that contains numeric codes instead of QOG value labels.
#' Datasets outside of the QOG Standard series are available only as Stata items
#' and require that \code{file} ends in \code{.dta} when \code{version} is not
#' \code{std}. The only exception is dataset \code{csyom}, which automatically
#' sets \code{version} to \code{std} and requires \code{file} to end
#' in \code{.csv}. Filenames with inadequate extensions will be modified to
#' conform to these expectations if they do not.
#' @seealso \code{\link{qogbook}}, \code{\link[foreign]{read.dta}}, \code{\link[foreign]{read.spss}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # Show URL to QOG Standard cross-section.
#' qogdata()
#' # Show URL to QOG Social Policy time series, long format.
#' qogdata(version = "soc", format = "tsl")
#' ## Download codebook and recent years from QOG Basic cross-section (not run).
#' # QOG = qogdata(file = "qog.cs.txt", version = "bas", format = "cs",
#' # years = 2002:2012, codebook = TRUE)
#' ## Download QOG Standard cross-section years of measurement (not run).
#' # QOG = qogdata(tempfile(fileext = ".csv"), format = "csyom")
#' ## Show QOG years of measurement for Gini coefficient (not run).
#' # table(QOG$wdi_gini)
#' @keywords qog
qogdata <- function(file = FALSE, replace = FALSE, codebook = FALSE, path = "",
version = "std", format = "cs",
variables = NULL, years = NULL, ...) {
try_require("foreign")
#
# currently available
#
versions = list(
std = c("ts", "cs", "csyom"),
bas = c("ts", "cs"),
exp = c("ctry", "ind"),
soc = c("cs", "tsl", "tsw"))
#
# correct version
#
if(!version %in% names(versions)) {
stop("Invalid version: use one of ",
paste0(names(versions), collapse = ", "))
}
#
# correct format
#
if(format == "csyom") {
version = "std"
if(!grepl(".csv$|.txt$", file))
file = gsub("(\\.|\\w){4}$", ".csv", file)
}
if(!format %in% unlist(versions[version])) {
stop("Invalid format: use one of ",
paste0(unlist(versions[version]), collapse = ", "))
}
#
# automatic filename
#
if(isTRUE(file)) {
file = paste0("qog_",
version,
"_",
format,
ifelse(version == "std",
paste0("_", "15May13.csv"),
".dta")
)
}
else {
if(is.character(file) & version != "std" & !grepl(".dta$", file)) {
file = gsub("(\\.|\\w){4}$", ".dta", file)
warning("QOG datasets other than std are available only as Stata files.\n",
" The filename that you specified was modified to ", file)
}
}
if(is.character(path))
if(nchar(path) > 0) file = paste(path, file, sep = "/")
#
# online source
#
link = paste0("http://www.qogdata.pol.gu.se/data/",
ifelse(version == "std", "QoG", "qog"),
"_", version, "_", format,
ifelse(version == "std", paste0("_", "15May13"), ""),
ifelse(version == "std" & grepl("csv|dta|sav", file),
substring(file, nchar(file) - 3),
".dta")
)
if(is.logical(file)) {
return(link)
}
else {
if(replace || !file.exists(file)) {
message("Downloading ", link, "...")
download.file(link, file, mode = "wb", quiet = TRUE)
}
else {
message("Loading from disk...")
}
}
#
# reader call
#
read = "read.csv"
args = list(file = file, ...)
# foreign or not
if(!grepl(".dta$|.sav$", file))
args["sep"] = ";"
# stata args
if(grepl(".dta$", file)) {
read = "read.dta"
if(is.null(unlist(args["warn.missing.labels"])))
args["warn.missing.labels"] = FALSE
}
# spss args
if(grepl(".sav$", file)) {
read = "read.spss"
if(is.null(unlist(args["to.data.frame"])))
args["to.data.frame"] = TRUE
}
data = do.call(read, args)
#
# selected variables
#
uids = c("ccode", "ccodealp", "cname", "ccodecow", "ccodewb")
pids = uids %in% names(data)
# avoid ts bug
if(grepl("ts|tsl", format) & !any(pids))
stop("You are trying to load a QOG dataset as time series, but it has no identifier variable.")
# avoid ts bug
if(grepl("ts|tsl", format) & !"year" %in% names(data))
stop("You are trying to load a QOG dataset as time series, but it has no year variable.")
if(!is.null(variables)) {
if(grepl("ts|tsl", format) & !"year" %in% variables) {
warning("Forcing year identifier into the dataset.")
variables = c("year", variables)
}
if(grepl("std|bas|soc", version) & !"ccode" %in% variables) {
warning("Forcing ccode identifier into the dataset.")
variables = c("ccode", variables)
}
data = data[, names(data) %in% variables]
}
#
# selected years
#
if(!is.null(years) && format %in% c("ts", "tsl", "ind"))
data = data[data$year %in% years, ]
#
# message
#
message("Loaded ", file, " (N = ", nrow(data),
ifelse(format %in% c("ts", "tsl", "ind"),
paste0(", ", min(data$year),
"-", max(data$year),
", T = ", length(unique(data$year))),
""),
").")
#
# grab codebook
#
if(isTRUE(codebook) || grepl(".pdf", codebook))
qogbook(codebook, version, path, replace)
#
# xtdata spec
#
pids = uids[pids]
if(format == "ts" | format == "tsl") {
data = xtset(data,
data = c(pids[1], "year", pids[-1]),
spec = c("iso3n", "year"),
type = "country",
name = "Quality of Government, time series data"
)
}
#
# finish line
#
return(data)
}
#' Download Quality of Government codebooks
#'
#' Function to download Quality of Government (QOG) codebooks. Please visit
#' the QOG Institute website at \url{http://www.qog.pol.gu.se/} for a
#' presentation of QOG research.
#'
#' @export
#' @param file a filename to save the codebook at.
#' If set to \code{TRUE}, the name of the codebook on the QOG server will be used.
#' If set to \code{FALSE} (the default), the function only returns the link to
#' the dataset. The filename must end in \code{.pdf}.
#' @param replace whether to download the dataset even if a file already exists
#' at the download location. Defaults to \code{FALSE}.
#' @param path a folder path to append to the filename.
#' @param version the QOG version: \code{std} (Standard),
#' \code{soc} (Social Policy), \code{bas} (Basic) or \code{exp} (Expert).
#' Defaults to \code{std}.
#' @details The function mimics Richard Svensson's \code{qogbook} Stata command.
#' @seealso \code{\link{qogdata}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # Show the URL to the QOG Standard dataset codebook.
#' qogbook()
#' ## Download QOG Standard codebook with default filename (not run).
#' # qogbook(file = TRUE)
#' ## Download QOG Basic dataset codebook to specific filename (not run).
#' # qogbook(file = "qog.basic.codebook.pdf", version = "bas")
#' @keywords qog
qogbook <- function(file = FALSE, version = "std", path = "", replace = FALSE) {
if(!version %in% c("std", "bas"))
stop("Codebook available only for versions bas, std")
link = "http://www.qogdata.pol.gu.se/data/Codebook_QoG_Std15May13.pdf"
if(version == "bas") {
link = "http://www.qogdata.pol.gu.se/codebook/codebook_basic_20120608.pdf"
}
if(isTRUE(file) & version == "bas") {
file = "codebook_basic_20120608.pdf"
}
else if(isTRUE(file)) {
file = "Codebook_QoG_Std15May13.pdf"
}
#
# path
#
if(is.character(path))
if(nchar(path) > 0) file = paste(path, file, sep = "/")
#
# download
#
if(is.logical(file)) {
return(link)
}
else if(!grepl(".pdf$", file)) {
stop("Please specify a .pdf codebook filename or set file = TRUE.")
}
else {
if(replace || !file.exists(file)) {
message("Downloading codebook to ", file, "...")
download.file(link, file, mode = "wb", quiet = TRUE)
}
message("Codebook: ", file)
}
}
#' Find Quality of Government variables
#'
#' Function to perform a \code{regex} search on QOG variable names and labels.
#' A few labels are missing for strictly cross-sectional variables.
#'
#' @export
#' @param ... keywords or \code{regex} phrases passed to \code{grepl}.
#' @param version the QOG version to search: either \code{std} (the default) or \code{soc}.
#' @param compact whether to limit the labels returned to 32 characters. Defaults to \code{TRUE} for better console output.
#' @param show which variables to show for years of measurement: \code{cs} (cross-sectional), \code{ts} (time series), or \code{all} (the default).
#' @return a data frame containg matching variables, described by their names, labels and years of measurement in the time series (\code{ts}) cross-sectional (\code{cs}) datasets. The information should match the ranges indicated in the \emph{QOG Standard Codebook} and \emph{QOG Social Policy Codebook}.
#' @references
#' Svensson, Richard, Stefan Dahlberg, Staffan Kumlin & Bo Rothstein.
#' 2012. \emph{The QoG Social Policy Dataset}, version 4Apr12.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#'
#' Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Soren Holmberg,
#' Bo Rothstein, Petrus Sundin & Richard Svensson. 2013.
#' \emph{The Quality of Government Dataset}, version 15May13.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#' @seealso \code{\link{grep}}, \code{\link{qogdata}}
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @examples
#' # QOG Standard search.
#' qogfind("regime", "institutions")
#' # QOG Standard search, with regex syntax.
#' qogfind("public|administration")
#' # QOG Social Policy search, showing cross-sectional information only.
#' head(qogfind("^socx", version = "soc", show = "cs", compact = FALSE))
#' # QOG Standard variables featured only in the cross-sectional version.
#' qogfind("*")[is.na(qogfind("*")$ts.N), ]
#' @keywords qog
qogfind <- function(..., version = "std", compact = TRUE, show = "all") {
x = paste0(c(...), collapse = "|")
data(qog.index)
if (version == "std") {
message("QOG Standard results")
r = qog.std.index[grepl(x, qog.std.index$variable, ignore.case = TRUE) |
grepl(x, qog.std.index$label, ignore.case = TRUE), ]
}
if (version == "soc") {
message("QOG Social Policy results")
r = qog.soc.index[grepl(x, qog.soc.index$variable, ignore.case = TRUE) |
grepl(x, qog.soc.index$label, ignore.case = TRUE), ]
}
if (version == "bas") {
message("QOG Basic results")
r = qog.bas.index[grepl(x, qog.bas.index$variable, ignore.case = TRUE) |
grepl(x, qog.bas.index$label, ignore.case = TRUE), ]
}
r$variable = as.character(r$variable)
if(compact) r$label = substr(r$label, 1, 32)
if(show == "cs") r = r[, !grepl("ts.", names(r))]
if(show == "ts") r = r[, !grepl("cs.", names(r))]
return(r)
}
#' Join historical and recent states in QOG Standard time series data
#'
#' Function to plot maps of Quality of Government (QOG) data. Requires the \code{ggplot2} and \code{maps} packages.
#'
#' @export
#' @param data a QOG Standard time series dataset, or any data frame with \code{cname} (country) and \code{year} information coded as in the QOG Standard time series dataset.
#' @param country the country name to join data over. Requires the \code{cname} variable.
#' @details The function will try to find two series of country-year observations that both match the \code{country} argument. Within the QOG Standard time series dataset, this will match historical states like "France (-1962)" to modern states like "France (1963-)". The function will then create a new variable out of both series, joined at their separation years, and set its country code attributes to the most recent ones. See Appendix A of the \emph{QOG Standard Codebook} for details on historical states in the QOG Standard time series dataset.
#' @return a data frame with country-year observations
#' @author Francois Briatte \email{f.briatte@@ed.ac.uk}
#' @references Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Soren Holmberg,
#' Bo Rothstein, Petrus Sundin & Richard Svensson. 2013.
#' \emph{The Quality of Government Dataset}, version 15May13.
#' University of Gothenburg: The Quality of Government Institute,
#' \url{http://www.qog.pol.gu.se}.
#' @examples
#' # Load QOG demo datasets.
#' data(qog.demo)
#' QOG = qog.ts.demo
#' QOG = qogjoin(QOG, "Ethiopia")
#' QOG = qogjoin(QOG, "France")
#' QOG = qogjoin(QOG, "Malaysia")
#' QOG = qogjoin(QOG, "Pakistan")
#' @keywords qog
qogjoin <- function(data, country = NULL) {
stopifnot("cname" %in% names(data))
stopifnot("year" %in% names(data))
stopifnot(is.character(country))
x = data$cname[grepl(country, data$cname)]
x = unique(x)
if(length(x) == 0)
stop("No country match for ", country)
else if(length(x) == 1)
stop("Single country match for ", country)
else if(length(x) > 2)
stop("More than two country matches: ", paste0(x, collapse = ", "))
t = xt(data)
y = as.numeric(gsub("\\D", "", x))
min = x[order(y) == 1]
max = x[order(y) == 2]
message("Joining ", min, " to ", max)
one = data[data$cname == min & data$year <= y[order(y) == 1], ]
two = data[data$cname == max & data$year >= y[order(y) == 2], ]
new = rbind(one, two)
new$cname = country
# country codes
if("ccode" %in% names(data))
new$ccode = unique(data$ccode[data$cname == max])
if("ccodealp" %in% names(data))
new$ccodealp = unique(data$ccodealp[data$cname == max])
if("cname_year" %in% names(data))
new$cname_year = paste(new$cname, new$year)
if("ccodealp_year" %in% names(data))
new$ccodealp_year = paste(new$ccodealp, new$year)
if("ccodewb" %in% names(data))
new$ccodewb = unique(data$ccodewb[data$cname == max])
message(country, " now runs over ", nrow(new), " country-year observations, ",
paste0(range(new$year), collapse = "-"))
data[, "cname"] = as.character(data[, "cname"])
data[data$cname == max, ] = new
data = subset(data, cname != min)
data[, "cname"] = factor(data[, "cname"])
data = xtset(data,
data = t$data,
spec = t$spec,
name = t$name,
url = t$url,
quiet = FALSE)
return(data)
}
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.