The Bernie Effect

Political piece here. Click to enlarge chart.



#Looking at net change, flux, other characteristics between voterdb pulls in Whatcom County.
# 12:12 PM 4/25/2016
library(data.table)
library(dplyr)
library(lubridate)
# utility Function
End <- function() {barplot(rep(1,8), yaxt="n", col=palette())}

vdb2014 <- fread("C:/Politics/11.22.2014/11.22.2014.txt")
vdb2015 <- fread("C:/Politics/RyanFerrisList_20151028/10.28.15.txt")
vdb2016 <- fread("C:/Politics/RyanFerris_20160415_094952_JessieS/RyanFerris_20160415_094952_JessieS.txt")
setkey(vdb2014,RegistrationNumber)
setkey(vdb2015,RegistrationNumber)
setkey(vdb2016,RegistrationNumber)

#Populate the years as appropriate with fun1()
#vdb <- vdb2014
#vdb <- vdb2015
#vdb <- vdb2016

fun1 <<- function()
{
vdb <- data.table(vdb)
vdb$RegistrationYear <<- with(vdb,year(mdy(RegistrationDate)))
vdb$LastUpdateYear <<- with(vdb,year(ymd_hms(LastUpdateDate)))
vdb$BirthYear <<- with(vdb,year(mdy(BirthDate)))
setkey(vdb,RegistrationNumber)

vdbA <<- data.table(vdb[StatusCode == "A",])
vdbA$RegistrationYear <<- with(vdbA,year(mdy(RegistrationDate)))
vdbA$LastUpdateYear <<- with(vdbA,year(ymd_hms(LastUpdateDate)))
vdbA$BirthYear <<- with(vdbA,year(mdy(BirthDate)))

vdbI <<- data.table(vdb[StatusCode == "I",])
vdbI$RegistrationYear <<- with(vdbI,year(mdy(RegistrationDate)))
vdbI$LastUpdateYear <<- with(vdbI,year(ymd_hms(LastUpdateDate)))
vdbI$BirthYear <<- with(vdbI,year(mdy(BirthDate)))
}

vdb <- vdb2016
fun1()
rY2016 <- arrange(count(vdb,RegistrationYear),desc(RegistrationYear))
lY2016 <- arrange(count(vdb,LastUpdateYear),desc(LastUpdateYear))
bY2016 <- arrange(count(vdb,BirthYear),desc(BirthYear))
rN2016 <- arrange(count(vdb,RegistrationNumber),desc(RegistrationNumber))
rNA2016 <- arrange(count(vdbA,RegistrationNumber),desc(RegistrationNumber))
rNI2016 <- arrange(count(vdbI,RegistrationNumber),desc(RegistrationNumber))

vdb <- vdb2015
fun1()
rY2015 <- arrange(count(vdb,RegistrationYear),desc(RegistrationYear))
lY2015 <- arrange(count(vdb,LastUpdateYear),desc(LastUpdateYear))
bY2015 <- arrange(count(vdb,BirthYear),desc(BirthYear))
rN2015 <- arrange(count(vdb,RegistrationNumber),desc(RegistrationNumber))
rNA2015 <- arrange(count(vdbA,RegistrationNumber),desc(RegistrationNumber))
rNI2015 <- arrange(count(vdbI,RegistrationNumber),desc(RegistrationNumber))

vdb <- vdb2014
fun1()
rY2014 <- arrange(count(vdb,RegistrationYear),desc(RegistrationYear))
lY2014 <- arrange(count(vdb,LastUpdateYear),desc(LastUpdateYear))
bY2014 <- arrange(count(vdb,BirthYear),desc(BirthYear))
rN2014 <- arrange(count(vdb,RegistrationNumber),desc(RegistrationNumber))
rNA2014 <- arrange(count(vdbA,RegistrationNumber),desc(RegistrationNumber))
rNI2014 <- arrange(count(vdbI,RegistrationNumber),desc(RegistrationNumber))

# Some Computations
# %in%
count(data.table(rNA2014$RegistrationNumber %in% rNA2015$RegistrationNumber),V1);
count(data.table(rNA2015$RegistrationNumber %in% rNA2016$RegistrationNumber),V1);
count(data.table(rNA2014$RegistrationNumber %in% rNA2016$RegistrationNumber),V1);

count(data.table(rNA2015$RegistrationNumber %in% rNA2014$RegistrationNumber),V1);
count(data.table(rNA2016$RegistrationNumber %in% rNA2015$RegistrationNumber),V1);
count(data.table(rNA2016$RegistrationNumber %in% rNA2014$RegistrationNumber),V1);

library(Hmisc)
# %nin%
count(data.table(rNA2014$RegistrationNumber %nin% rNA2015$RegistrationNumber),V1);
count(data.table(rNA2015$RegistrationNumber %nin% rNA2016$RegistrationNumber),V1);
count(data.table(rNA2014$RegistrationNumber %nin% rNA2016$RegistrationNumber),V1);

count(data.table(rNA2015$RegistrationNumber %in% rNA2014$RegistrationNumber),V1);
count(data.table(rNA2016$RegistrationNumber %in% rNA2015$RegistrationNumber),V1);
count(data.table(rNA2016$RegistrationNumber %in% rNA2014$RegistrationNumber),V1);

t1 <- vdb2016[StatusCode == "A",]
t2 <- vdb2014[StatusCode == "A",]
new <- t1[RegistrationNumber %nin% t2$RegistrationNumber,.(RegistrationNumber,BirthDate,PrecinctID,RegistrationDate)]
old <- t2[RegistrationNumber %nin% t1$RegistrationNumber,.(RegistrationNumber,BirthDate,PrecinctID,RegistrationDate)]

BirthYear <- aggregate(RegistrationNumber ~ year(mdy(BirthDate)),data=new,length)
colnames(BirthYear) <- c("BirthYear", "Count")
newAge <- arrange(BirthYear,desc(BirthYear))

PrecinctID <- aggregate(RegistrationNumber ~ PrecinctID,data=new,length)
colnames(PrecinctID) <- c("PrecinctID", "Count")
newPrecinct <- arrange(PrecinctID,desc(PrecinctID))

newAge <-  data.table(newAge)
newPrecinct <-  data.table(newPrecinct)
N <- nrow(new)
P <- newPrecinct[PrecinctID > 199 & PrecinctID < 300,sum(Count)]
A <- newAge[BirthYear > 1966,sum(Count)]
newbd <- with(new,year(mdy(BirthDate)))
oldbd <- with(old,year(mdy(BirthDate)))

# Some Charts
par(mfrow=c(1,2))
plot(newAge,col="blue")
plot(newPrecinct,col="red")
par(mfrow=c(1,1))
mtext(paste0(prettyNum(N,big.mark=","), " new Whatcom County Registrants by Birth Year and PrecinctID added between 11/22/2014 and 4/15/2016."),cex=1.25, line = 1)
mtext(paste0(prettyNum(P,big.mark=","), " new registrants come from the 200 (Bellingham) series. ",prettyNum(A,big.mark=","), " new registrants are less than 50 years old (e.g BirthYear > 1966)."),cex=1.25,line = 0)

par(mfrow=c(1,2))
plot(density((oldbd)),xlim=c(1900,2010),ylim=c(0,.06),lwd=2,col="blue")
plot(density((newbd)),xlim=c(1900,2010),ylim=c(0,.06),lwd=2,col="red")
par(mfrow=c(1,1))
End()

# Complicated to interpret but interesting
par(mfrow=c(1,3))
plot(bY2016,type="l",col="red1");lines(bY2015,col="blue2");lines(bY2014,col="green3")
plot(lY2016,type="l",col="red1");lines(lY2015,col="blue2");lines(lY2014,col="green3")
plot(rY2016,type="l",col="red1");lines(rY2015,col="blue2");lines(rY2014,col="green3")
par(mfrow=c(1,1))

par(mfrow=c(1,3))
plot(rN2016[,RegistrationNumber,],type="l",col="red1");lines(rN2015[,RegistrationNumber,],col="blue2");lines(rN2014[,RegistrationNumber,],col="green3")
plot(rNA2016[,RegistrationNumber,],type="l",col="red1");lines(rNA2015[,RegistrationNumber,],col="blue2");lines(rNA2014[,RegistrationNumber,],col="green3")
plot(rNI2016[,RegistrationNumber,],type="l",col="red1");lines(rNI2015[,RegistrationNumber,],col="blue2");lines(rNI2014[,RegistrationNumber,],col="green3")
par(mfrow=c(1,1))
End()