################################################################################################
##########################   EstaTiDados    ####################################################
########################## GUSTAVO BRUSSE   ####################################################
################################################################################################

# Baixando pacotes
install.packages("devtools")
install.packages("stringi") 
devtools::install_github("lucasmation/microdadosBrasil")

# Libraries
library('microdadosBrasil')
library(dplyr)
library(data.table)
library(ggplot2)
require(gridExtra)

# Diretorio origem
setwd("D:\\Gustavo\\Zjg1ZGMzMmRlOThiNGRjZD\\Volume{970c5b77-5f9c-4b5d-af34-3a3b0c6b725e}\\Users\\Gustavo\\Desktop\\Doutorado\\PROJETO EDSD")

# Download PNAD 2011
download_sourceData("PNAD", 2011, unzip = T)

# Lendo banco de Domicilios
d  <- read_PNAD("domicilios", 2011)

# Lendo banco de Pessoas
p <- read_PNAD("pessoas", 2011)

# Selecionando apenas variaveis: estado (V0102), variavel controle (V0103), sexo (V0302), idade (V8005)
# relacao com pessoa de referencia (V0401)
data.1<-p[,c(1, 2, 4, 8, 9)]

# Selecionando alguns estados SP (35), RJ (33), MG (31) e ES (32)
data.1$V0102<-as.numeric(data.1$V0102)
data.2<-subset(data.1, V0102>=34999999 & V0102<=35999999 | 
                 V0102>=32999999 & V0102<=33999999 |
                 V0102>=30999999 & V0102<=31999999 |
                 V0102>=32000000 & V0102<=32999999)

# Deletando missing values
data.3<-subset(data.2, complete.cases(data.2)==TRUE)

# Criando variavel controle unica
data.3$id <- as.numeric(paste(data.3$V0102, data.3$V0103, sep = ""))


####### algoritmo para criar tipos de domicilios ##########################################################3

# Identificando pessoas de referencia
data.3$V0401<-as.numeric(data.3$V0401)
data.3$head<-as.numeric(data.3$V0401==1)

# Identificando conjuges
data.3$spouse<-as.numeric(data.3$V0401==2)

# Identificando filhos e enteados
data.3$son<-as.numeric(data.3$V0401==3)

# Identificando outros parentes
data.3$other<-as.numeric(data.3$V0401==4)

# Identificando nao parentes
data.3$non.relatives<-as.numeric(data.3$V0401>4)

# Agregando individuos em domicilios usando a variavel controle 
data.4<-data.3[,c( 11, 6, 7, 8, 9, 10)]
data.4$id<-as.numeric(data.4$id)
data.5<-aggregate(data.4, by=list(data.4$id), FUN=sum)

# Criando a variavel "tipo de domicilio" (HH)
data.5$HH<-NA

# Tipo 1: Unipessoal
data.5$HH[data.5$spouse==0 & data.5$son==0 & data.5$other==0 & data.5$non.relatives==0]<-1

# Tipo 2: Casal sem filhos + outros
data.5$HH[data.5$spouse==1 & data.5$son==0 & data.5$non.relatives==0]<-2

# Tipo 3: Monoparental + outros
data.5$HH[data.5$spouse==0 & data.5$son>0 & data.5$non.relatives==0]<-3

# Tipo 4: Casal com filhos + outros
data.5$HH[data.5$spouse==1 & data.5$son>0 & data.5$non.relatives==0]<-4

# Tipo 5: Estendidos ou compostos
data.5$HH[data.5$non.relatives>0 | data.5$spouse==0 & data.5$son==0 & data.5$other>0]<-5

# renomear categorias
levels(data.5$HH) <- c("unipessoal", "casal_sem_filhos", "monoparental", "casal_com_filhos", "estendido_composto" )

# checar missing values
sum(is.na(data.5$HH))

# Tabela de frequencia
t<-table(data.5$HH)
addmargins(t) 

## Fazendo merge da variavel 'tipo de domicilio' na base dos individuos
data.6<-data.5[,c( 1, 8)]
names(data.6)<-c('id', 'HH')

data.7<-merge(data.3, data.6, by = "id")
data.8<-data.7[,c(4, 5, 7, 12)]
names(data.8)<-c('sex','age','head','HH')

########### Calculando Taxas de Chefia #################################### 

# Criando variavel grupo de idade
data.8$age<-as.numeric(data.8$age)
data.8$ageclass<-NA
data.8$ageclass[data.8$age>=0 & data.8$age<=4]<-1
data.8$ageclass[data.8$age>=5 & data.8$age<=9]<-2
data.8$ageclass[data.8$age>=10 & data.8$age<=14]<-3
data.8$ageclass[data.8$age>=15 & data.8$age<=19]<-4
data.8$ageclass[data.8$age>=20 & data.8$age<=24]<-5
data.8$ageclass[data.8$age>=25 & data.8$age<=29]<-6
data.8$ageclass[data.8$age>=30 & data.8$age<=34]<-7
data.8$ageclass[data.8$age>=35 & data.8$age<=39]<-8
data.8$ageclass[data.8$age>=40 & data.8$age<=44]<-9
data.8$ageclass[data.8$age>=45 & data.8$age<=49]<-10
data.8$ageclass[data.8$age>=50 & data.8$age<=54]<-11
data.8$ageclass[data.8$age>=55 & data.8$age<=59]<-12
data.8$ageclass[data.8$age>=60 & data.8$age<=64]<-13
data.8$ageclass[data.8$age>=65 & data.8$age<=69]<-14
data.8$ageclass[data.8$age>=70 & data.8$age<=74]<-15
data.8$ageclass[data.8$age>=75 & data.8$age<=79]<-16
data.8$ageclass[data.8$age>=80]<-17

# Variavel sexo como numerico
data.8$sex<-as.numeric(data.8$sex)
str(data.8)

# Tabulando os dados por sexo e grupo de idade
tab<-table(data.8$HH,data.8$sex,data.8$ageclass)   
tab1<-table(data.8$HH,data.8$sex,data.8$ageclass,data.8$head)   

# funcao que calcula a Taxa de Chefia
pop<-vector(l=170)
head<-vector(l=170)
pos=1
for(i in 1:5)
{
  for(j in 1:2)
  {
    for(k in 1:17)
    {
      
      pop[pos]<-sum(tab[,j,k])
      head[pos]<-tab1[i,j,k,2]
      pos<-pos+1
      
    }
  }
}

df<-data.frame(HH=c(rep(1,34),rep(2,34),rep(3,34),rep(4,34),rep(5,34)),
               sex=rep(c(rep(1,17),rep(2,17)),5),
               age=rep(1:17,10),
               head=head,
               pop=pop)

## Checando os resultados
tab[5,2,17]  # pop
tab1[5,2,17,2]  # head

# Criando variavel Taxa de Chefia
df$HSrate2001<-round(df$head/df$pop,4)

# Saving database in .CSV
write.csv(df, file = "2011.csv")

######  Plotting  ########
# Male - Casal sem filhos
maleCNC<-subset(df, df$sex==1 & df$HH==2)
plot_maleCNC<-ggplot(maleCNC, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.45)+
  labs(title="Homens - Casal sem filhos",x="Grupo de idade", y = "População")
plot_maleCNC

# Female - Couple without (no) children
femaleCNC<-subset(df, df$sex==2 & df$HH==2)
plot_femaleCNC<-ggplot(femaleCNC, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.5)+
  labs(title="Female - Couple without (no) children",x="Age groups", y = "Population")
plot_femaleCNC

# Male - Lone Parente
maleLP<-subset(df, df$sex==1 & df$HH==3)
plot_maleLP<-ggplot(maleLP, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.6)+
  labs(title="Male - Lone Parente",x="Age groups", y = "Population")
plot_maleLP

# Female - Lone Parente
femaleLP<-subset(df, df$sex==2 & df$HH==3)
plot_femaleLP<-ggplot(femaleLP, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.6)+
  labs(title="Female - Lone Parente",x="Age groups", y = "Population")
plot_femaleLP

# Male - Couple with children
maleCWC<-subset(df, df$sex==1 & df$HH==4)
plot_maleCWC<-ggplot(maleCWC, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.5)+
  labs(title="Male - Couple with children",x="Age groups", y = "Population")
plot_maleCWC

# Female - Couple with children
femaleCWC<-subset(df, df$sex==2 & df$HH==4)
plot_femaleCWC<-ggplot(femaleCWC, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.5)+
  labs(title="Female - Couple with children",x="Age groups", y = "Population")
plot_femaleCWC

# Male - All others
maleAO<-subset(df, df$sex==1 & df$HH==5)
plot_maleAO<-ggplot(maleAO, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.5)+
  labs(title="Male - All others",x="Age groups", y = "Population")
plot_maleAO

# Male - All others
femaleAO<-subset(df, df$sex==2 & df$HH==5)
plot_femaleAO<-ggplot(femaleAO, aes(x=age, y=HSrate2001)) +
  geom_line(size=1.7,color="grey69")+
  geom_smooth(span = 0.5)+
  labs(title="Male - All others",x="Age groups", y = "Population")
plot_femaleAO

grid.arrange(plot_maleCNC, plot_femaleCNC, plot_maleLP, plot_femaleLP, plot_maleCWC,
             plot_femaleCWC, plot_maleAO, plot_femaleAO, ncol=2)