-
Notifications
You must be signed in to change notification settings - Fork 0
/
Further Data Manipulation.R
31 lines (26 loc) · 1.25 KB
/
Further Data Manipulation.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
#################################################
## further data manipulation ##
#################################################
# create a data matrix
dataMatrix <- cbind(dataMatrix, VAN.ID = df$VAN.ID, County.Voter.ID = df$County.Voter.ID)
dataMatrix <- cbind(dataMatrix, outcome = df$outcome)
# get a list of precinct data frames
designMatrices <- lapply(trainPrecincts, FUN = function(p) {
dataMatrix[df$Precinct == p,]
})
names(designMatrices) <- trainPrecincts
names(designMatrices)[names(designMatrices) == "Wharton Boro 05"] <- "Wharton Borough D05"
# clean things up
trainMatrix <- do.call(rbind, designMatrices)
outcome <- trainMatrix[,ncol(trainMatrix)]
trainMatrix <- trainMatrix[,-ncol(trainMatrix)] # drop the outcome
IDs <- trainMatrix[,c("VAN.ID", "County.Voter.ID")]
trainMatrix <- trainMatrix[,-((ncol(trainMatrix)-1):ncol(trainMatrix))] # drop the IDs
trainMatrix <- data.frame(trainMatrix)
designMatrices <- lapply(designMatrices, FUN = function(x) {
x[,!colnames(x) %in% c("VAN.ID", "County.Voter.ID")]
})
# make a holdout matrix
holdoutMatrix <- dataMatrix[df$Precinct %in% holdoutPrecincts,]
holdoutMatrix <- holdoutMatrix[,!colnames(holdoutMatrix) %in%
c("VAN.ID", "County.Voter.ID")]