Inferring gender from column of first names in R

gistfile1.txt

# Let's do some advanced stuff. 
# First we have to isolate the first name from the NAME field

payroll$first_name <- gsub(".*\\,", "", payroll$NAME)
payroll$first_name <- gsub(" .*", "", payroll$first_name)

# Bring in a library to help normalize the cases
library(stringr)
payroll$first_name <- str_to_title(payroll$first_name)

# Inferring gender from first name based on historical data from the Census and Social Security admin
install.packages("gender")
library(gender)

# Running the gender function on the payroll column that has the first names
# This will probably take several minutes to process. Have patience. 
payroll_gender<- gender(payroll$first_name)

# You'll now have a 9mb list. Convert it to a data frame to merge with the original payroll set
payroll_gender <- do.call(rbind, lapply(payroll_gender, data.frame, stringsAsFactors=FALSE))


# Isolating the new gender dataframe to just name and gender
payroll_gender <- payroll_gender[c("name", "gender")]

# Cleaning it up to match with the original data frame to make joining easier
colnames(payroll_gender) <- c("first_name", "gender")

# Deleting duplicates. Makes it easier to join.
payroll_gender <- subset(payroll_gender, !duplicated(payroll_gender$first_name))

# Bringing in a package that helps with advanced data work
library(dplyr)

# Bringing together the payroll data set with the inferred gender data set
payroll <- left_join(payroll, payroll_gender)

# If no names were matched with a gender, set it to "unknown
payroll$gender[is.na(payroll$gender)] <- "unknown"

# Let's take a look at the breakdown
table(payroll$gender)