I can’t install Ibgepesq in R to read the Pnads

Question

I just downloaded the 2013 PNAD of IBGE -- and I tried to open it with the package for R that IBGE itself elaborated, the Ibgepesq. It is available as . zip archive at this address:

ftp://ftp.ibge.gov.br/Trabalho_e_Rendimento/Pesquisa_Nacional_por_Amostra_de_Domicilios_anual/microdados/2013/Leitura_em_R.zip

I downloaded it into my Working directory. And ran:

    install.packages("IBGEPesq_1.0-4.zip",
                     repos=NULL)

    library(IBGEPesq)

But then I get the following message:

    Error: package ‘IBGEPesq’ was built before R 3.0.0: please re-install it

Obviously, I’ve tried to re-install it. And I’ve also run it here:

    # Para atualizar os demais pacotes
    update.packages(checkBuilt = TRUE, ask = FALSE)

    # Por recomendação em um fórum (não entendi bem por quê) 
    install.packages('codetools')

And even then, the re-installation does not work. The same error returns. I am using a Windows 7 and the R session data is:

    > sessionInfo()
    R version 3.1.1 (2014-07-10)
    Platform: x86_64-w64-mingw32/x64 (64-bit)

    locale:
    [1] LC_COLLATE=Portuguese_Brazil.1252  LC_CTYPE=Portuguese_Brazil.1252   
    [3] LC_MONETARY=Portuguese_Brazil.1252 LC_NUMERIC=C                      
    [5] LC_TIME=Portuguese_Brazil.1252    

    attached base packages:
    [1] stats     graphics  grDevices utils     datasets  methods   base     

    loaded via a namespace (and not attached):
    [1] tools_3.1.1

Beforehand, I appreciate the help.

Answer 1

I recommend using the script created by Damico and Djalma, which can be found here. I’ve worked with him a lot and it works perfectly.

You get the data via: download all Microdata. R, and then perform the analysis with single-year - analysis examples. R.

There you will also find scripts for SME, POF and PISA.

Answer 2

I know Damico’s scripts, but personally I prefer the solution I’m going to present here. Because the PNAD data is provided as microdata, it is sufficient to have the research dictionary that the reading is trivial using any reader with delimiter. For performance reasons I will use data.table here and a C++ function of the desc package that converts delimited text files into csv files, fwf2csv(). Then just use the fread() function of the data.table that reads csv’s super fast.

Initially you will need the dictionary and microdata, both of which can be downloaded here: http://www.ibge.gov.br/home/estatistica/populacao/trabalhoerendimento/pnad2012/microdados.shtm In the spreadsheet "Dictionary of Household Variables of the Basic Research - 2013.xls" export to a CSV the first three columns: Initial Position, Size, Variable Code. In my case I saved as dicdom.csv.

With the microdata in the Data folder run the script:

#############PREPARAÇÃO DE DADOS##########
library(bit64)
library(data.table)
library(descr)
library(xlsx)

## Criando o dicionário a partir das três primeiras colunas da planilha
dicdom <- read.csv(file = 'dicdom.csv', header=F)
dicdom <- dicdom[complete.cases(dicdom),]
colnames(dicdom) <- c('inicio', 'tamanho', 'variavel')

## Parâmetro com o final de cada campo
end_dom = dicdom$inicio + dicdom$tamanho - 1

## Converte o microdado para um arquivo csv
fwf2csv(fwffile='Dados/DOM2013.txt', csvfile='dadosdom.csv', names=dicdom$variavel, begin=dicdom$inicio, end=end_dom)

## Efetua a leitura do conjunto de dados com o fread do data.table
dadosdom <- fread(input='dadosdom.csv', sep='auto', sep2='auto', integer64='double')

And that’s it! There are 148,697 households, just check with nrow(). Repeat the procedure for people’s data.