Judicial Analytics

An Exploration of the Influence of Partisan Politics on the US Circuit Judges from 1927 to 2010

Carp-Manning U.S. District Court Database

Robert A. Carp and Kenneth L. Manning. 2016. “U.S. District Court Database.” 2016 version n=110977. URL: http://districtcourtdatabase.org

Team challenges: We have very limited coding experience.

Team accomplishments: We improved our comfort levels in R and Python by diving into the analysis of an unknown dataset.

R code

mydata <- hackathon #csv file

install.packages("remotes")

library('devtools') install_github('brooksandrew/Rsenal')

library('Rsenal')

install.packages("devtools")

remotes::install_github("brooksandrew/Rsenal")

??Rsenal

install.packages("ggplot2")

library(ggplot2) library(gridExtra) library(grid) library(cluster) library(data.table)

ave_dec <- aggregate(mydata[, 9:10], list(mydata$year), mean)

st.err <- function(x) { sd(x)/sqrt(length(x)) }

ste <- aggregate(mydata[, 9:10], list(mydata$year), st.err)

ave_dec

dec_yr_cir <- aggregate(mydata[, 9:10], list(mydata$year), mean)

dec_yr_cir1 <- aggregate(mydata[, 9:10], list(mydata$year, mydata$circuit), mean)

dec_yr_cir_typ <- aggregate(mydata[, 9:10], list(mydata$year, mydata$circuit, mydata$casetype), mean)

dec_yr_cir_typ

dec_yr_cir_gen <- aggregate(mydata[, 9:10], list(mydata$year, mydata$circuit, mydata$gender), mean)

nrow(dec_yr_cir1)

lvl <- levels(as.factor(mydata$year)) mydata <- cbind(mydata, "avg" = rep(NA, nrow(mydata))) for (i in 1:nrow(mydata)) { for ( l in lvl ) { if (mydata$year[i] == l) { mydata$avg[i] <- ave_dec$libcon[which(ave_dec$year == l)] }}} times <- c(1927, 2010) subst_times <- NULL for ( i in 1:nrow(mydata)) { if ( mydata$year[i] > times[1] & mydata$year[i] < times[2] ){ subst_times <- rbind(subst_times, mydata[i,]) } } breaks <- NULL i <- 1928 while ( i <= 2010){ breaks <- c(breaks, i) i <- i + 4 }

lvl <- levels(as.factor(mydata$year))

mydata <- cbind(mydata, "ste" = rep(NA, nrow(mydata)))

for (i in 1:nrow(mydata)) {

for ( l in lvl ) {

if (mydata$year[i] == l) {

mydata$avg[i] <- ste$libcon[which(ste$year == l)]

}}}

times <- c(1927, 2010)

subst_times <- NULL

for ( i in 1:nrow(mydata)) {

if ( mydata$year[i] > times[1] & mydata$year[i] < times[2] ){

subst_times <- rbind(subst_times, mydata[i,])

}

breaks <- NULL

i <- 1928

while ( i <= 2010){

breaks <- c(breaks, i)

i <- i + 4

}

mydata <- mydata[,-c(20:21)]

ggplot() + geom_point(data=mydata, mapping=aes(x = mydata$year, y = mydata$avg)) + theme_classic() + geom_hline(yintercept = .5, color = "red") #+ geom_vline(xintercept = breaks)

mydata2 <- mydata2[,-c(20:21)] ggplot() + geom_point(data=mydata2, mapping=aes(x = mydata2$year, y = mydata2$avg)) + theme_classic() + geom_hline(yintercept = .5, color = "red") #+ geom_vline(xintercept = breaks)

ggplot() + geom_smooth(data=dec_yr_cir1, mapping=aes(x = dec_yr_cir1$year, y = dec_yr_cir1$libcon, colour=factor(dec_yr_cir1$Group.2)), se = F) + theme_classic() + geom_hline(yintercept = .5, color = "red")

ggplot() + geom_smooth(data=dec_yr_cir1, mapping=aes(x = dec_yr_cir1$year, y = dec_yr_cir1$libcon), se = T) + theme_classic() + geom_hline(yintercept = .5, color = "red") + facet_grid(rows=vars(Group.2))

ggplot() + geom_smooth(data=dec_yr_cir_gen, mapping=aes(x = dec_yr_cir_gen$year, y = dec_yr_cir_gen$libcon, colour=factor(dec_yr_cir_gen$Group.3)), se = T) + theme_classic() + geom_hline(yintercept = .5, color = "red") #keep female vs male US judges, period where female judges were significantly

dec_party_typ <- aggregate(mydata[, 10:11], list(mydata$casetype, mydata$party), mean)

dec_party_typ