Package 'MaddisonData'

Get Maddison sources

Description

The Maddison project collates historical economic statistics from many sources.

They have a citation policy: CONDITIONS UNDER WHICH ALL ORIGINAL PAPERS MUST BE CITED:

a) If the data is shown in any graphical form b) If subsets of the full dataset that include less than a dozen (12) countries are used for statistical analysis or any other purposes

When neither a) or b) apply, then the MDP as a whole can be cited.

getMaddisonSources returns a data.frame of relevant sources for a particular application.

Usage

getMaddisonSources(
  ISO = NULL,
  plot = TRUE,
  sources = MaddisonData::MaddisonSources,
  years = MaddisonData::MaddisonYears
)

Arguments

ISO	either NULL to return all sources or a character vector of ISO codes for the countries included in the analysis or a data.frame with the first column being the ISO codes followed by yearBegin and optionally yearEnd.
plot	logical indicating whether the use does nor does not include plotting data. The Maddison project requires citing all relevant MaddisonSources if they are plotted, denoted here by plot = TRUE. If no data are plotted, then the Maddison project requires citing all sources only if less than a dozen are used, denoted here by plot = FALSE, in which case, the Maddison project requires a specific project-level citation. Default = TRUE.
sources	list of sources in the format of MaddisonSources; default is MaddisonSources.
years	data.frame in the format of MaddisonYears; default is MaddisonYears.

Value

a data.frame with 3 columns:

ISO

3-letter ISO code for country.

years

character vector of years or year ranges for which source applies.

source

character vector of sources.

in the format of MaddisonSources.

Examples

getMaddisonSources() # all 
getMaddisonSources(plot=FALSE) # only MDP 
GBR <- getMaddisonSources('GBR') # GBR 

getMaddisonSources(names(MaddisonSources)[1:12], FALSE) # only MDP 
getMaddisonSources(data.frame(ISO=c('GBR', 'USA'), 
             yearBegin=rep(1500, 2)) ) #GBR, USA since 1500 
getMaddisonSources('AUS') # AUS: no special sources for AUS.

---

ggplot paths

Description

ggplotPath plots y vs. x (typically year) with a separate line for each group with options for legend placement, horizontal and vertical lines and labels.

Usage

ggplotPath(
  x = "year",
  y,
  group,
  data,
  scaley = 1,
  logy = TRUE,
  ylab,
  legend.position,
  hlines,
  vlines,
  labels,
  fontsize = 10,
  color,
  linetype
)

Arguments

x	name of a numeric column in data to pass as x in aes(x=.data[[x]], ...); default = year.
y	name of column in data to pass as y in aes(y=.data[[y]], ...); must be supplied.
group	name of grouping variable, i.e., plot a separate line for each level of group using aes(group=.data[[group]], ...), unless group is missing or length(unique(data[, group])) = 1.
data	data.frame or tibble::tibble with columns x, y, and group.
scaley	number to divide y by for plotting. Default = 1, but for data in monetary terms, e.g., for MaddisonData, y = 'gdppc' is Gross domestic product (GDP) per capita in 2011 dollars at purchasing power parity (PPP), for which we typically want scaley = 1000.
logy	logical: if TRUE, y axis is on a log scale; default = TRUE.'
ylab	y axis label. Default = if(scaley==1) y else paste(y, '/', scaley)
legend.position	argument passed to ggplot2::theme. If !missing(labels), default is no legend. Otherwise, default depends on ⁠nGps <- length(unique(data[, group]⁠: If nGps = 1, there is no legend. If ⁠nGps < 10, legend.position = c(.15, .5)⁠. Else if nGps < 2-. legend.position = 'right'. Else legend.position = 'none'. To suppress the legend, use legend.position = 'none'. For other alternatives, see ggplot2::theme.
hlines	numeric vector of locations on the y axis for horizontal lines using ggplot2::geom_hline(yintercept = hlines, ...) with ⁠color='grey', lty='dotted'⁠ unless color or colour and / or lty are available as attr(x, ...).
vlines	numeric vector of locations on the x axis for vertical lines using ggplot2::geom_vline(xintercept = vlines, ...) with ⁠color='grey', lty='dotted'⁠ unless color or colour and / or lty are available as attr(x, ...).
labels	= data.frame with columns ⁠x, y, label⁠, and optionally ⁠srt, col, size⁠, where x, y, srt, and size are are numeric, label is character, and col are acceptable values for color in with(labels, annotate('text', x=x, y=y, label = label, srt=srt, color=col, size=size)). Defaults for srt, col, and size are 0, 'black', and 4, respectively.
fontsize	for legend and axes labels in theme(text=element_text(size=fontsize)); default = 10.
color	for lines to pass to scale_color_manual(values = color) if present. If present, length(color) should equal length(unique(data[, group])).
linetype	optional vector. Default ⁠if(missing(group))⁠ 1 else ⁠rep(1:6, length=length(unique(data[, group]))⁠. Else either an integer (0-6), a name (0 = blank, 1 = solid, 2 = dashed, 3 = dotted, 4 = dotdash, 5 = longdash, 6 = twodash), or a mapping to a discrete variable, or a string of an even number (up to eight) of hexadecimal digits which give the lengths in consecutive positions in the string.

Value

an object of class ggplot2::ggplot, which can be subsequently edited, and whose print method produces the desired plot.

Examples

str(GBR_USA <- subset(MaddisonData::MaddisonData, ISO %in% c('GBR', 'USA')))
GBR_USA1 <- ggplotPath('year', 'gdppc', 'ISO', GBR_USA, 1000)
GBR_USA1a <- ggplotPath('year', 'gdppc', 'ISO', GBR_USA, 1000, 
                        color=c('red', 'blue'))

GBR_USA1+ggplot2::coord_cartesian(xlim=c(1500, 1850)) # for only 1500-1850 
GBR_USA1+ggplot2::coord_cartesian(xlim=c(1600, 1700), ylim=c(.9, 3)) 

# label the lines
ISOll <- data.frame(x=c(1500, 1800), y=c(2.5, 1.7), label=c('GBR', 'USA'), 
              srt=c(0, 30), col=c('red', 'green'), size=c(2, 9))
GBR_USA2 <- ggplotPath('year', 'gdppc', 'ISO', GBR_USA, 1000, 
                    labels=ISOll, fontsize = 20)  
                        
# h, vlines, manual legend only 
Hlines <- c(1,3, 10, 30)
Vlines = c(1849, 1929, 1933, 1939, 1945)
(GBR_USA3 <- ggplotPath('year', 'gdppc', 'ISO', GBR_USA, 1000, 
       ylab='GDP per capita (2011 PPP K$)', 
       legend.position = NULL, hlines=Hlines, vlines=Vlines, labels=ISOll))  

# do.call(ggplotPath, ...) with 1 line
list1 <- list(x='Time', y='lvl', data=data.frame(Time=1:4, lvl=sqrt(1:4)))
doCallPlot <- do.call(ggplotPath, list1)

---

ggplotPath in multiple panels with no space between panels and a shared horizontal axis on the bottom.

Description

ggplotPath2 accepts a variety of inputs. The most general is with object = a list of sublists to be fed individually to do.call(ggplotPath,..) and then assembled into the desired plot after eliminating the space between the individual plots. Optional ⁠\dots⁠ arguments give default values for the individual calls to ggplotPath.

Usage

ggplotPath2(object, ...)

## S3 method for class 'list'
ggplotPath2(object, ...)

## S3 method for class 'mts'
ggplotPath2(object, Time, ...)

## S3 method for class 'KFS'
ggplotPath2(object, ...)

## Default S3 method:
ggplotPath2(
  object,
  Time,
  object2,
  scaley,
  logy,
  ylab,
  hlines,
  vlines = numeric(0),
  labels,
  fontsize = 10,
  color,
  linetype,
  ...
)

Arguments

object	with an associated matrix to be plotted. The following classes of objects are supported: matrix, data.frame, tibble or mts of dim = c(n, k), e.g., k components of a state vector at n points in time. ggplotPath is called separately for each of the k column. Each such call returns a ggplot object. Those ggplot objects are assembled as separate panels into a ggplot object with a shared horizontal axis with no space between the panels with at most one more row than length(Time) and the first row of object is dropped if needed so n == length(k). KFS (defined in package KFAS) calls ggplotPath.matrix(object$a, ...). list with a component a, assumed to a model produced from estimation in the KFAS package, calls ggplotPath2(object$a, ...). list with component model: If it has a component a, call ggplotPath2(object$model$a, ...). Else call ggplotPath2(KFAS::KFS(object), ...).
...	optional arguments.
Time	optional numeric vector with lenght being either the number of rows of the matrix obtained from object or one less and fed to ggplotPath to create the individual panels of the plot. Default depends on class(object): matrix: Default Time = rownames(object). mts: Default Time = time(object). KFS (defined in package KFAS): Default = rownames or time of object$model$y. model (defined in package KFAS): Default = rownames or time of object$y.
object2	an optional vector or matrix-type object with the same number of rows as object or one less and k2 columns that provide separate reference lines to appear in the same panels as the first k2 columns of object. Obviously, k2 <= k. (Also, legend.position = 'none'.)
scaley	= optional numeric vector of the length k fed individually as the scaley argument accompanying the different successive calls to ggplotPath. If the matrix obtained from object has two columns with names = c('level', 'slope') or c('level', 'growthRate'), then the default for scaley = c(1000, .0). Otherwise, default = 1.
logy	optional character vector of length k to control the scales used in the individual panels of the plot: exp_log: Feed exp of column i of the matrix to plot to ggplotPath(..., logy=TRUE). log: Feed column i of the matrix to plot to ggplotPath(..., logy=TRUE). ”: Feed column i of the matrix to plot to ggplotPath(..., logy=FALSE). The default is 'exp_log' for the first panel and ” for the rest.
ylab	optional character vector of length k for y axis labels.
hlines	optional list of at most k of numeric vectors (possibly with attributes), for horizontal lines in panel i, feeding hlines[[i]] to ggplotPath(..., hlines = hlines[[i]]) for panel i = 1:k. ⁠color='grey', lty='dotted'⁠ unless color or colour and / or lty are available as attr(hlines[[i]], ...). hlines outside the plot range are suppressed with a warning.
vlines	optional numeric vector of locations on the x axis for vertical lines using ggplotPath(..., vlines=vlines) with ⁠color='grey', lty='dotted'⁠ unless color or colour and / or lty are available as attr(vlines, ...). vlines outside the plot range are suppressed with a warning.
labels	= optional data.frame with columns ⁠x, y, label, component⁠, and optionally ⁠srt, col, size⁠, where x, y, srt, and size are are numeric, label is character, and col are acceptable values for color in with(labels, annotate('text', x=x, y=y, label = label, srt=srt, color=col, size=size)). Defaults for srt, col, and size are 0, 'black', and 4, respectively. component is an integer in 1:k. labelsi = subset(labels, component==i) is used in ggplotPath(..., labels=labelsi) to label panel i = 1:k. labels outside the plot range are suppressed with a warning. labels near the plot boundaries may be clipped.
fontsize	optional number for legend and axes labels, used in 'ggplotPath(..., fontsize=fontsize); default = 10.
color	optional vector or list of length 'k of vectors to feed to ggplotPath(..., color=color[[i]]) for panel i=1:k. If present, length(color) should equal k. Default is c('black', 'red') for each panel with a reference line provided in object2 and black otherwise. See ggplotPath for options.
linetype	optional vector or list of length k of vectors to feed to ggplotPath(..., linetype=linetype[[i]]) for panel i=1:k. If present, length(linetype) should equal k. Default for each panel with a reference line in object2 = c(1, 3) = c('solid', 'dotted') else 1 = solid. See ggplotPath for options.

Details

Alternatively, the first object argument can be a matrix, data.frame, tibble, or multivariate time series (of class mts). An optional Time argument must be a numeric vector with length equal to the number of rows of object. If Time is provided, it overrides any values for the horizontal axis that could be inferred from rownames or time of object. A second optional object2 can be either a vector whose length matches the number of rows of object or a matrix or ts object similarly matching the number of rows of object.

Value

an object of class ggplot2::ggplot, which can be subsequently edited, and whose print method produces the desired plot.

Examples

# matrix examples 
Mat <- cbind(lvl=1:5, vel=rep(1:2, length=5), acc=sin(1:5))
Mat1 <- Mat
rownames(Mat1) <- 1951:1955
Mat2 <- as.data.frame(cbind(Mat, year=1951:1955))

# mts example 
MTS <- ts(Mat, 1951)

# Do 
Matp <- ggplotPath2(Mat)
# with object2 = vector 
  Matp1 <- ggplotPath2(Mat, object2=sqrt(1:5))
# with object2= = 2 column matrix for first 2 panels. 
Matp2 <- ggplotPath2(Mat, object2=2*Mat[, 2:1])
Mat1p <- ggplotPath2(Mat1)
Mat2p <- ggplotPath2(Mat2[, 1:3], Time=Mat2[, 'year']) 
MTSp <- ggplotPath2(MTS)
MTSep <- ggplotPath2(MTS, logy=c('exp_log', 'log', ''))

# list example 
List2 <- list(
    level=list('year', 'lvl'), 
    slope=list('year', 'vel'), 
    accel=list('year', 'acc'))
Mat2l <- ggplotPath2(List2, data=Mat2)

# State space / Kalman filtering model for GBR
GBR <- subset(MaddisonData, (ISO=='GBR') & !is.na(gdppc))
# model example 
growthFormula <- (log(gdppc)~ -1 + SSMbespoke(growthModel(.04, GBR$gdppc) )) 
library(KFAS)
GBR2m <-SSModel(growthFormula, GBR, H=matrix(NA) )

# NOTE: This call ignores Time 
GBRgrowthFit1 <- fitSSM(GBR2m, inits=-6, method = "BFGS", 
                        updatefn = growthUpdateFn)
# NOTE: This call currently also ignores Time; MUST BE FIXED
GBRgrowthFit1t <- fitSSM(GBR2m, inits=-6, method = "BFGS", 
                         updatefn = growthUpdateFn, Time=GBR$year)
GBRfitp <- ggplotPath2(GBRgrowthFit1t)

#KFS example 
GBR_KFS <- KFAS::KFS(GBRgrowthFit1$model)
GBR_KFSt <- KFAS::KFS(GBRgrowthFit1t$model)
GBR_KFSp0 <- ggplotPath2(GBR_KFS)
GBR_KFSp <- ggplotPath2(GBR_KFS$a)
GBR_KFStp <- ggplotPath2(GBR_KFSt$a)

# label the lines
ISOll1 <- data.frame(x=c(1500, 1800), y=c(2.5, 1.7), 
                 label=c('GBR', 'Napoleon'), srt=c(0, 30),
                 col=c('red', 'green'), size=c(2, 9), 
                 component=1)

GBR_KFSp1 <- ggplotPath2(GBR_KFS$a, labels=ISOll1)
                 
GBR_KFSp <- ggplotPath2(GBR_KFS, labels=ISOll1)
ISOll2 <- ISOll1
ISOll2$component <- 1:2
GBR_KFSp2 <- ggplotPath2(GBR_KFS, labels=ISOll2)

# hlines, vlines 
zero <- 0
attr(zero, 'color') <- 'red'
attr(zero, 'lty') <- 'dashed'
Hlines1 <- list(c(1,3, 10, 30), zero) 
Vlines <- c(1649, 1929, 1933, 1945)

GBR_KFSp3 <- ggplotPath2(GBR_KFS, labels=ISOll2, hlines=Hlines1, 
                         vlines=Vlines)

---

Two-dimensional growth model for KFAS

Description

growthModel returns a list returned by KFAS::SSMcustom() for a model with potentially irregularly spaced univariate observations with a 2-dimensional (level, growthRate) state.

Usage

growthModel(
  sigma,
  y,
  a1,
  Time,
  stateNames = c("level", "growthRate"),
  Log = TRUE,
  ...
)

Arguments

sigma	a numeric vector, forced to length 2 by replacing it by rep(sigma, length=2). The one-step transition variance for level is Q[1,1] = sigma[1]^2 and for growthRate is Q[2,2] = sigma[2]^2. With m missing values, Q[1,1] = (m+1)*sigma[1]^2, Q[2,2] = (m+1)*sigma[2]^2, and ⁠Q[1,2] = Q[2,1] = sqrt((m+1)*choose(m+1,2)*sigma[1]*sigma[2]⁠.
y	= optional numeric vector or ts object of length at least 2 used to estimate the a1 parameter for KFAS::SSMcustom() if a1 is not supplied and for Time if it is not supplied. a1 <- (if(Log) log(c(y[1], y[2]/y[1])) else c(y[1], y[2]-y[1])).
a1	= optional numeric vector of length 2 to pass to KFAS::SSMcustom. If supplied, the numeric values of y are ignored.
Time	= optional integer vector of times at which non-missing observations are available. Default = time(y) if y is of class ts or or names(y) if !is.null or 1:length(y) otherwise.
stateNames	= c('level', 'growthRate')
Log	default = TRUE.
...	optional arguments passed to KFAS::SSMcustom().

Value

a list returned by KFAS::SSMcustom() with an additional Time component.

Examples

GBR <- subset(MaddisonData, (ISO=='GBR') & !is.na(gdppc))
growthMdl1 <- growthModel(.1, GBR$gdppc, Time=GBR$year)
GBRgdppc1 <- with(GBR[-1, ], ts(gdppc, year[1]))
growthMdl2 <- growthModel(c(.1, .2), GBRgdppc1)
growthMdl0 <- growthModel(.1, GBR$gdppc, a1=c(10, 1), Log=FALSE, 
                  stateNames=c('lvl', 'vel'))

---

Update function to estimating a two-dimensional growth model

Description

growthUpdateFn

Usage

growthUpdateFn(pars, model, Time)

Arguments

pars	= log(variance) of noise for ⁠[1] observations⁠, ⁠[2] level⁠, and ⁠[3] growthRate⁠. Forced to length 3 via c(pars, rep(tail(pars, 1), length=3-length(pars))). With no missing values, the one-step transition variance for level is ⁠Q[1,1] = exp(pars[2]⁠ and for growthRate is Q[2,2, ] = exp(pars[3]). With m missing values, Q[1,1, ] = (m+1)*exp(pars[2]), Q[2,2] = (m+1)*exp(pars[3]), and ⁠Q[1,2, ] = Q[2,1] = sqrt((m+1)*choose(m+1,2)*exp(mean(pars[2:3]))⁠.
model	= list assumed to have components n = third dimension of the transition covariance array Q and state_names = c('level', 'growthRate') as returned by growthModel.
Time	= optional integer vector of times at which non-missing observations are available. Default is model$Time if available or time(y) if model has a component y or names(y) if !is.null or 1:n otherwise.

Details

returns a list returned by KFAS::SSMcustom() for a model with potentially irregularly spaced univariate observations with a 2-dimensional (level, growthRate) state.

Value

a model with components H and Q updated as described.

Examples

GBR <- subset(MaddisonData, (ISO=='GBR') & !is.na(gdppc))
growthMdl1 <- growthModel(.1, GBR$gdppc, Time=GBR$year)
growthMdl1v0 <- growthUpdateFn(.1, growthMdl1)
growthMdl1v <- growthUpdateFn(.1, growthMdl1, Time=GBR$year)
growthMdl1v3 <- growthUpdateFn(1:3, growthMdl1, Time=GBR$year)

growthFml <- (log(gdppc)~ -1 + SSMbespoke(growthModel(.1, GBR$gdppc) )) 
library(KFAS)
growthSSmdl <-SSModel(growthFml, GBR, H=matrix(NA) )
growthSSmdl1 <- growthUpdateFn(.1, growthSSmdl)

---

Select countries and add logged variables

Description

logMaddison returns a tibble::tibble of data on selected countries extracted from MaddisonData, appending columns lnGDPpc and lnPop = natural logarithms of gdppc and pop.

Usage

logMaddison(ISO = NULL)

Arguments

ISO	either NULL to select all the data in MaddisonData or a character vector of ISO codes used in the Maddison project.

Value

a tibble::tibble with 6 columns:

ISO

3-letter ISO code for countries selected

year

numeric year in the current era.

gdppc

Gross domestic product per capita adjusted for inflation to 2011 dollars at purchasing power parity.

pop

Population, mid-year (thousands)

lnGDPpc

log(gdppc)

lnPop

log(pop)

Examples

logMaddison() # all 
logMaddison(c('GBR', 'USA')) # GBR, USA

---

Convert a vector of date ranges into a data.frame

Description

MadDateRanges returns a data.frame with 3 numeric columns: yearBegin, yearEnd, and sourceNum from the vector of dateRanges associated with different sources in MaddisonSources.

Usage

MadDateRanges(dateRanges)

Arguments

dateRanges

character vector of date ranges, each associated with a different source.

Value

a data.frame with 3 columns

yearBegin, yearEnd

numeric years

sourceNum

1, 2, 3, ... for the location in dateRanges

Examples

MadDateRanges(c('1', '700 – 1500', '1252–1700 (England)', 
      '1915-1919 & 1949', '1820, 1870, 1913, 1950'))
# equal 
data.frame(
yearBegin=c(1,  700, 1252, 1820, 1870, 1913, 1950), 
yearEnd  =c(1, 1500, 1700, 1820, 1870, 1913, 1950), 
sourceNum=c(1, 2, 3, rep(4, 4)))

---

Maddison Project data

Description

The Maddison project collates historical economic statistics from many sources. MaddisonCountries is a data.frame of all (countrycode, country, region) combinations in those data.

Usage

MaddisonCountries

Format

MaddisonCountries

A data frame with 3 columns:

ISO

3-letter ISO country code

country

Country name used by the Maddison project

region

Geographic region including country

Its rownames = ISO.

Source

https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2020?lang=en"Groningen Growth and Development Centre"

Examples

# Get the country for a countrycode (IS)
subset(MaddisonCountries, ISO=='GBR', country)
# Or
MaddisonCountries['GBR', 'country']
# Find Yugoslavia 
subset(MaddisonCountries, grepl('Yugo', country), 1:3)
# number of countries by region 
table(MaddisonCountries$region)
# What are "Western Offshoots"? 
subset(MaddisonCountries, grepl('Of', region), c(country, ISO))

---

Maddison Project data

Description

The Maddison project collates historical economic statistics from many sources. MaddisonCountries is a data.frame of all (countrycode, country, region) combinations in those data. This object provides easy access to the 2023 version of the Maddison project data downloaded 2025-08-28.

Usage

MaddisonData

Format

MaddisonData

A data frame with 4 columns:

ISO

3-letter ISO country code

year

numeric year starting with year 1 CE

gdppc

Gross domestic product (GDP) per capita in 2011 dollars at purchasing power parity (PPP)

pop

Population, mid-year (thousands)

Source

https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2020?lang=en"Groningen Growth and Development Centre"

Examples

# Get the countrycode for a country
subset(MaddisonCountries, country=='United Kingdom', ISO)
# Select  
str(GBR <- MaddisonData[MaddisonData$ISO=='GBR', ])

---

Identify leading countries

Description

MaddisonLeaders computes the countries with the highest gdppc for each year.

Usage

MaddisonLeaders(
  except = character(0),
  y = "gdppc",
  group = "ISO",
  data = MaddisonData::MaddisonData,
  x = "year"
)

Arguments

except	either NULL to select all the data in MaddisonData or a character vector of group codes to EXCLUDE, e.g., so the result reflects apparent technology leaders, excluding countries whose high gdppc may be due to a dominant position in a single commodity.
y	name of column in data to consider. Default = gdppc.
group	name of column in data as the grouping variable. Default = ISO.
data	data.frame or tibble::tibble with first two columns being ISO and year and y being the name of another column.
x	time variable. Default = year.

Value

an object of class c('MaddisonLeaders', 'data.frame'), with columns

• ⁠paste0(x, 'Begin)⁠,

• paste0(x, 'End'),

• paste0(y, '0'),

• paste0(y, '1'), and

• {{group}}

• ⁠paste0('d', x, '0') = ⁠ paste0(x, 'End') - paste0(x, 'Begin') + min(dx), where dx = min(diff(sort(unique(data[, x]))))

• ⁠paste0('d', x, '1') = ⁠ c(tail(paste0(x, 'Begin'), -1) - head(paste0(x, 'End'), -1), NA) (defaults: dy0 = yearEnd - yearBegin +1 and dy1 = c(tail(yearBegin, -1) - head(yearEnd, -1), NA) )

(defaults:

• yearBegin,

• yearEnd,

• gdppc0,

• gdppc1, and

• ISO, plus

• dyear0 = yearEnd - yearBegin + 1 and

• dyear1 = c(tail(yearBegin, -1) - head(yearEnd, -1), NA)

with an attribute LeaderByYear = a data.frame with columns, {{x}}, paste0('max', y), and {{group}} (defaults: year, maxgdppc, ISO).

Examples

Leaders0 <- MaddisonLeaders() # max GDPpc for each year. 

# Presumed technology leaders without commodity leaders with narrow 
# economies 
Leaders1 <- MaddisonLeaders(c('ARE', 'KWT', 'QAT')) 

# since 1600 
MadDat1600 <- subset(MaddisonData, year>1600)
Leaders1600 <- MaddisonLeaders(c('ARE', 'KWT', 'QAT'), data=MadDat1600)

# max pop by region within percentiles of gdppc
noGDP <- is.na(MaddisonData$gdppc)
MadDat <-MaddisonData[!noGDP, ]
gdpPcts <- quantile(MadDat$gdppc, seq(0, 1, .01), na.rm=TRUE)
gdpPct <- unique(as.numeric(gdpPcts[-1]))
gdpPc <-c(gdpPct[-100], tail(gdpPct, 1)*(1+sqrt(.Machine$double.eps)))
gdp100 <- MadDat$gdppc
nObs <- nrow(MadDat)
for(i in 1:nObs){gdp100[i] <- min(gdpPc[MadDat$gdppc[i]<gdpPc])}
MadDat$gdp100 <- gdp100
MadDat$region <- MaddisonCountries[MadDat$ISO, 'region', drop=TRUE]
MadPopRgnGDP<-MaddisonLeaders(y='pop',group='region',data=MadDat,x='gdp100')

---

Maddison Project data

Description

The Maddison project collates historical economic statistics from many sources. MaddisonSources is a list of tibble::tibbles with ISO names giving the sources of GDP per capita for different years for the said country.

MaddisonYears is a data.frame giving yearBegin and yearEnd and the number of each source in MaddisonSpources for each ISO.

Usage

MaddisonSources

MaddisonYears

Format

MaddisonSources

A named list of tibble::tibbles, one for each country, named with the ISO country codes. Each tibble has one row for each source for the indicated ISO and two columns:

years

character variable of year(s) for this source starting with year 1 CE.

source

character variable giving the source for the years described.

In addition, MaddisonSources has an attribute since2008, which says, "gdppc since 2008: Total Economy Database (TED) from the Conference Board for all countries included in TED and UN national accounts statistics for all others."

MaddisonYears

A data.frames with 4 columns:

ISO

3-letter country code.

yearBegin, yearEnd

Integer year begin and end for each source.

sourceNum

Integer of the source within MaddisonSources[[ISO]].

An object of class data.frame with 133 rows and 4 columns.

Source

https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2020?lang=en"Groningen Growth and Development Centre"

Examples

MaddisonSources[['GBR']]
MaddisonSources[['GBR']][, 1, drop=TRUE] 
# = c('1', '1252–1700 (England)', '1700–1870') 
# for data from the year 1 
# and for England only between 1252 and 1700, etc. 

MaddisonSources[['IRN']][, 1, drop=TRUE] 
# = '1820, 1870, 1913, 1950'
# for those 4 years only. 

MaddisonSources[c('GBR', 'USA')]

MaddisonSources[['GBR']][, 1, drop=TRUE] 
# = c('1', '1252–1700 (England)', '1700–1870') 

MaddisonYears[MaddisonYears$ISO=='GBR', ] = 
data.frame(
ISO=rep('GBR', 3), 
yearBegin=c(1, 1252, 1700), 
yearEnd  =c(1, 1700, 1870), 
sourceNum=1:3
)

MaddisonSources[['EGY']][, 1, drop=TRUE] 
# = c('1', '700 – 1500', '1820, 1870, 1913, 1950')

MaddisonYears[MaddisonYears$ISO=='EGY', ] = 
data.frame(
ISO=rep('EGY', 6), 
yearBegin=c(1,  700, 1820, 1870, 1913, 1950), 
yearEnd  =c(1, 1500, 1820, 1870, 1913, 1950), 
sourceNum=c(1,    2, rep(3, 4))
)

---

Construct a path to a location within an installed or development package

Description

path_package2 returns a character vector of matches to target. It differs from system.file() in that it supports searching for a target file or folder possibly in subdirs of the working directory or in nparents of its parents.

Usage

path_package2(
  target,
  package = NULL,
  nparents = 1,
  subdirs = c("extdata", paste("inst", "extdata", sep = .Platform$file.sep))
)

Arguments

target	A regular expression describing the file of folder desired.
package	Name of the package to in which to search. If NULL, search in the working directory. Otherwise search in system.file(package).
nparents	integer indicate the number of parents of the working directory in which to search; default = 1.
subdirs	= c('extdata', paste('inst','extdata', sep=.Platform$file.sep))

Details

This works in a vignette searching for a target that could be in the vignettes directory of its parent package or in the package directory or in, e.g., one of subdirs = c('extdata', paste('inst', 'extdata', sep=.Platform$file.sep)).

Returns the full path to match(s) if found and a character vector of length 0 if no matches are found. The returned object also has a searched attribute being a character vector of the directories searched.

This was inspired by a desire to share with others a vignette describing how to create data objects from a file that could not itself be shared on CRAN. This is not easy, because the working director available to code in a vignette changes depending on how that code is run.

path_package2 allows the user to store the target locally, e.g., in inst/extdata but include it in .gitignore to prevent it from leaving the local computer. The vignette then decides what to do after calling path_package2() based on the length of the the object returned.

Value

a character vector with an attribute searched giving the full paths of all directories searched for target.

Examples

# search for a file matching a regular expression
path_package2('^mpd.*xlsx$') 
# search only in the working directory
path_package2('^mpd.*xlsx$', nparents=0, subdirs=character(0))

---

Summary method for an object of class MaddisonLeaders

Description

summary.MaddisonLeaders returns a data.frame with columns ISO, ⁠paste0(x, 'Begin)⁠, paste0(x, 'End'), n, and p.

Usage

## S3 method for class 'MaddisonLeaders'
summary(object, sortBy = "ISO", decreasing = FALSE, ...)

Arguments

object	= object of class MaddisonLeaders.
sortBy	= column of output used for sorting; default = ISO
decreasing	default = FALSE
...	= optional arguments for summary (not used)

Value

a data.frame with columns

• ISO = One row for each level of ISO in unique(object[, 'ISO'])

• ⁠paste0(x, 'Begin)⁠ = earliest object[, paste0(x, 'Begin')] for ISO

• paste0(x, 'End'), last object[, paste0(x, 'End')] for ISO

• n = sum of ⁠(paste0(x, 'End')⁠ - paste0(x, 'Begin') + 1 for ISO.

• p = n/(paste0(x, 'End') - paste0(x, 'Begin') + 1).

)

(defaults:

• ISO = One row for each level of ISO in unique(object[, 'ISO'])

• yearBegin = earliest ⁠object[, 'yearBegin')]⁠ for ISO

• yearEnd = last ⁠object[, 'yearEnd')]⁠ for ISO

• n = sum of ('yearEnd' - 'yearBegin' + 1) for ISO.

• p = n/(yearEnd - yearBegin + 1).

[, 'yearBegin')]: R:,%20'yearBegin') [, 'yearEnd')]: R:,%20'yearEnd')

Examples

Leaders0 <- MaddisonLeaders() # max GDPpc for each year. 
summary(Leaders0)

---

year with fraction'

Description

yr converts a Date to a year and fraction. For example, 2025-01-01 becomes 2025.00000, while 2025-01-02 becomes 2025.00234, because (2-1)/365 is 0.00234 to 5 significant digits. However, 2024-01-02 becomes 2024.0233, because (2-1)/366 is only 0.00233 to 5 significant digits.

Usage

yr(x, ...)

Arguments

x	quantity that can be converted to a Date object using as.Date(x).
...	arguments passed to lubridate::ymd().

Value

a number (numeric vector).

See Also

lubridate::decimal_date(), lubridate::ymd()

Examples

Jan2_24_25 <- c('2024-01-02', '2025-01-02')
J2yr <- yr(Jan2_24_25)
J2y <- yr(as.POSIXct(Jan2_24_25))
all.equal(J2yr, J2y)

---