Social-security-Disability-case-study-

Social security Disability case study

Load in the libraries that we'll need

pacman::p_load(tidyverse, lubridate, stringr)

loading the dataset

social_security <- read_csv("ssadisability.csv")

glimpse(social_security) view(social_security) attach(social_security)

this data seems so wide so we can make it longer so we get a single data in each row

social_security_longer <- pivot_longer(social_security, !Fiscal_Year, names_to = "month", values_to = "applications") view(social_security_longer)

Formatting the dates, we look at the unique values of the months

unique(social_security_longer$month)

this seems like there is a problem here as the internet application and the

normal application are all combined together, we need to separate them

social_security_longer <- social_security_longer %>% separate(month, c("month", "application_method"), sep = "_") view(social_security_longer)

we now have to look at the unique variables in the month

unique(social_security_longer$month)

you realised that "May" "June" "July" and "August" are written in full and this

might cause a problem so we have to extract only the first three letter of all the

month

social_security_longer <- social_security_longer %>% mutate(month = substr(month, start = 1, stop = 3)) view(social_security_longer)

Lets also have a look at the unique values of the fiscal years

unique(social_security_longer$Fiscal_Year)

this looks likes the years are not in the standard formats so we will have to replace

all the FYs with 20

social_security_longer <-social_security_longer %>% mutate(Fiscal_Year = str_replace(Fiscal_Year, pattern = "FY", replacement = "20")) view(social_security_longer)

Because the month and year are in a separate forms, we can combine them to be in

in a single from using the paste function

social_security_longer$date <- dmy(paste('01', social_security_longer$month, social_security_longer$Fiscal_Year)) view(social_security_longer)

this gives you a date to work with

Working with fiscal year and calender year, we need to convert the fiscal year to a calender year

social_security_longer <- social_security_longer %>% mutate(Fiscal_Year=as.numeric(Fiscal_Year)) %>% mutate(Fiscal_Year=ifelse(month(date)>=10, Fiscal_Year-1, Fiscal_Year)) %>% mutate(date=dmy(paste("01", month, Fiscal_Year)))

view(social_security_longer)

Changing the name well

social_security_longer <- social_security_longer %>% rename(Fiscal_Year = Fiscal_Year, Month = month, Application_methods = application_method, Applications = applications, Date = date)

view(social_security_longer) summary(social_security_longer) attach(social_security_longer)

We realized that the Fiscal year and the months are no more important so we get rid of them

also the Application method show that it is a character variable so we change it to factors

social_security_longer <- social_security_longer %>% select(-Fiscal_Year, -Month) %>% mutate(Application_methods = as.factor(Application_methods)) view(social_security_longer) summary(social_security_longer)

everything looks perfect now

social_security <- pivot_wider(social_security_longer, names_from = Application_methods, values_from = Applications) %>% select(Date, Internet, Total) attach(social_security) view(social_security)

Looking at the percentage of people who applied online

social_security <- social_security %>% mutate(Online_Percentage = Internet / Total * 100)

view(social_security)

Creating a scatter plot to see if internet application has grown over time

ggplot(data = social_security, mapping = aes(x = Date, y = Online_Percentage)) + geom_point()