weighByMultiClasses: Iterate calcShares over multiple groups

In mbannert/gateveys: An R Toolbox for Business Tendency Survey Researchers

Description

Iterate calcShares over multiple groups

Usage

  weighByMultiClasses(dtable, keyvector, variables, weight)

Arguments

dt	input data.table containing id,group and participant answer information
keyvector	vector containing the categories to group by
variables	vector of categorical variable, typically questions in survey
weight	vector containing quantitative weights, typically number of employees

Value

list of lists that contains results with groups on the toplevel and variables on the level velow

Author(s)

Matthias Bannert

See Also

capWeight,calcShares,burstList

Examples

# draw some weights from a mixed distribution
w <- drawFromMixed(1000,list(rnorm=list(mean=1000,sd=250),
                             rexp=list(),
                             rchisq=list(df=10)),
                             c(0.1,0.5,0.4))
w <- ceiling(unlist(w))
# create a basic categorical samle panel
sData <- generateSamplePanel(80,3,"q",c(1995,1),c(2001,3),weight=w)
# turn it into data.table to make use of indexing !
sData[,grep("quest",names(sData))] <- lapply(sData[,grep("quest",names(sData))],as.factor)

# add a NAs and group to the mix
sData <- generateRandomNAs(sData,c(1:3,7),30,1)
sData <- merge(sData,generateRandomGroups(3,unique(sData$uid)),by="uid")

# set some size class, only needed for the weighByMultiClasses function
sData <- setSizeClass(sData,thresholdList=list(M=20,L=200),size="weight",
             sectorColumn="all")

# turn data.frame to data.table to use the keys
sData.dt <- data.table(sData,key=c("year","period","group"))
# run calcShares
calcShares(sData.dt,"question_1","weight")

# grep questions
questions <- grep("^quest",names(sData),value=TRUE)

# apply weighByNOGA over all questions
listOfResults <- weighByMultiClasses(sData.dt,"group",questions,"weight")
names(listOfResults) <- "group"
# note that if group is character vector > 1 this lists contains multiple 
# elements each representing one class (aggregation level)

mbannert/gateveys documentation built on May 21, 2019, 2:23 p.m.