scripts/exec.dataset.construction.R

In jcizel/SovereignCrisis: What the package does (one line)

'require('WorldBankAPI')'
# ---------------------------------------------------------------------------- #
# NEED THE FOLLOWING MACROECONOMIC VARIABLES:                                  #
#                                                                              #
# 1. GOVERNMENT REVENUE                                                        #
# 2. GOVERNMENT EXPENDITURE                                                    #
# 3. EXTERNAL GOVERNMENT DEBT                                                  #
# 4. GDP                                                                       #
# 5. UNEMPLOYMENT                                                              #
# 6.                                                                           #
#                                                                              #
# ---------------------------------------------------------------------------- #

## GOVERNMENT REVENUE
options(width = 200)
queryMacroDatasets('balance of payments')
queryMacroDatasets('liquidity')
queryMacroDatasets('stock market')
queryMacroDatasets('performing')
queryMacroDatasets('tax')
queryMacroDatasets('crisis')
queryMacroDatasets('revenue')
queryMacroDatasets('expenditure')
queryMacroDatasets('debt') 
queryMacroDatasets('external')
queryMacroDatasets('tax revenue')
queryMacroDatasets('house')
queryMacroDatasets('gross domestic product')
queryMacroDatasets('gdp ')
queryMacroDatasets('gross national income')
queryMacroDatasets('unemployment')
queryMacroDatasets('employ')
queryMacroDatasets('treasury')
queryMacroDatasets('yield')
queryMacroDatasets('spread')
queryMacroDatasets('distance')
queryMacroDatasets('probability')
queryMacroDatasets('score')
queryMacroDatasets('arrear')
queryMacroDatasets('repayment')
queryMacroDatasets('ppg ')
queryMacroDatasets('disbursement')
queryMacroDatasets('interest payment')
queryMacroDatasets('interest rate')
queryMacroDatasets('default')
queryMacroDatasets('failure')
queryMacroDatasets('bankrupt')
queryMacroDatasets('short.term')
queryMacroDatasets('imf')
queryMacroDatasets('capital')
queryMacroDatasets('foreign direct investment')
queryMacroDatasets('polit')
queryMacroDatasets('growth')
queryMacroDatasets('index')
queryMacroDatasets('trade')
queryMacroDatasets('export')
queryMacroDatasets('import')
queryMacroDatasets('embi')
queryMacroDatasets('output gap')
queryMacroDatasets('official')
queryMacroDatasets('official credit')
queryMacroDatasets('terms')
queryMacroDatasets('terms.of.trade')
queryMacroDatasets('ted spread')
queryMacroDatasets('balance')
queryMacroDatasets('public sector debt')
queryMacroDatasets('GFDD')
queryMacroDatasets('budget')
queryMacroDatasets('deficit')
queryMacroDatasets('Government Finance')
queryMacroDatasets('corporate')
queryMacroDatasets('bail')

## IMF Lending




## MANUALLY MARK THE CSV FILES CREATED IN THE ABOVE QUERIES, BY ADDING A
## `SELECT` THAT EQUALS 1 IF THE VARIABLE IS TO BE INCLUDED IN THE FINAL DATASET. 

## -------------------------------------------------------------------------- ##
## CREATE ANALYSIS DATASET                                                    ##
## -------------------------------------------------------------------------- ##
crisis1 <- loadCrisisDB()
crisis2 <- alternativeCrisisDB()
alt     <- getAltmanZscore()
pd      <- getAggregatedBankscopePDs()
macro   <- createQueriedMacroDataset(test = FALSE)
macro %>>% save(file = './inst/extdata/rdata/macro-data.RData')

oecd.debt <- fread('../SDMXWrappers/inst/extdata/OECD/GOV_DEBT/DATATABLE.csv')
oecd.bank <- fread('../SDMXWrappers/inst/extdata/OECD/BPF1/DATATABLE.csv')



dtList <-
    list(
        crisis1,
        crisis2,
        alt,
        pd,
        macro,
        oecd.debt,
        oecd.bank
    )


dt <- dtList %>>% joinDatasetList()




dt <- augmentBenchmarkDataset(
    crisisdb = crisis1,
    dtList =
        list(crisis2,
             alt,
             pd,
             macro
             )
)




lookup[varcode %like% "CN$"][,list(varcode,label)]

dt %>>%
within({
    GDP     = NY.GDP.MKTP.CN    
    C.GDP.CN     = NY.GDP.MKTP.CN
    C.TAXREV.GDP = GC.TAX.TOTL.CN / C.GDP.CN
    C.INT.GDP    = GC.XPN.INTP.CN / C.GDP.CN
    C.INT.TAXREV = GC.XPN.INTP.CN / GC.TAX.TOTL.CN
}) -> dtmod

dtmod %>>%
(~ hist(.$C.TAXREV.GDP,100)) %>>%
(~ hist(.$C.INT.GDP,100)) %>>%
(~ hist(.$C.INT.TAXREV,100))


procExpand(
    data = dtmod,
    by = 'iso3',
    keepvars = c('date','GDP'),
    convert =
        list('
__CN$ ~ `/`(shift(GDP,lag = -1));
winsorize(method = "IQR")'),
    suffix = '.RATIO')


macro
%>>%
subset(date > '1960-01-01') %>>%
(? str(.)) %>>%
(df ~ df[, hist(DP.DOD.DECD.CR.GG.CD
               ,100)])



.cumprod <- function(x){
    x[is.na(x)] <- 1
    cumprod(x)
}

procExpand(
    data = dtmod,
    by = 'iso3',
    keepvars = c('date','GDP'),
    convert =
        ## list('GDP ~ shift(lag = -1, dif = TRUE, relative = TRUE)')        
        list('__CN$ ~ `/`(shift(GDP,lag = -1));winsorize(method = "IQR")')
)

dtmod %>>%
procExpand(by = 'iso3',
           keepvars = 'date',
           convert =
           list('L62 ~ shift(lag = -1, dif = TRUE, relative = TRUE)')) %>>%
subset(iso3=='USA' & date > '1960-01-01') %>>%
qplot(x = date,
      y = L62,
      data = .,
      geom = 'line')

lookup <- attributes(macro)$lookup

convertToStata(data = dt,
               lookup = lookup)