library(readxl)
library(ggplot2)
library(lubridate)
UnemploymentRate <- read_excel("C:/Users/shahi/Desktop/DAT 400/UNRATE.xls")
str(UnemploymentRate) ## Checking the data types for my variables
## Classes 'tbl_df', 'tbl' and 'data.frame': 860 obs. of 2 variables:
## $ observation_date: POSIXct, format: "1948-01-01" "1948-02-01" ...
## $ UNRATE : num 3.4 3.8 4 3.9 3.5 3.6 3.6 3.9 3.8 3.7 ...
colnames(UnemploymentRate)[colnames(UnemploymentRate)== "observation_date"] <- "Date" ##Renaming a column for easier coding later on
UnemploymentRate$Date <- as.Date(UnemploymentRate$Date) ## Changed this variable to data type Date
ggplot(data = UnemploymentRate, aes(x = Date, y = UNRATE))+
########This section of code adds the recession shading, has to be before the plot so it does not cover the plotted lines #################################
geom_rect(aes(xmin = ymd('1948-11-01'),
xmax = ymd('1949-10-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1953-07-01'),
xmax = ymd('1954-05-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1957-08-01'),
xmax = ymd('1958-04-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1960-04-01'),
xmax = ymd('1961-02-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1969-12-01'),
xmax = ymd('1970-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1973-11-01'),
xmax = ymd('1975-03-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1980-02-01'),
xmax = ymd('1980-07-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1981-07-01'),
xmax = ymd('1982-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1990-07-01'),
xmax = ymd('1991-03-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('2001-03-01'),
xmax = ymd('2001-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('2008-12-01'),
xmax = ymd('2009-06-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_line(color = "blue") +
########This Section of code changes the tick start date on the x & y axis ##################################################
scale_x_date( date_labels = ( "%Y"), breaks= as.Date(c("1950-01-01", "1955-01-01","1960-01-01","1965-01-01","1970-01-01","1975-01-01","1980-01-01","1985-01-01","1990-01-01","1995-01-01","2000-01-01","2005-01-01","2010-01-01","2015-01-01"))) +
scale_y_continuous(breaks = (1:11)) +
##### This section of code adds the proper titles #####################################
labs(
x = "Date",
y = "Percent",
title = "Civilian Unemployment Rate",
subtitle = "Source: https://fred.stlouisfed.org"
)
library(readxl)
library(ggplot2)
library(lubridate)
library(dplyr)
library(plyr)
library(sqldf)
UnemploymentRate <- read_excel("C:/Users/shahi/Desktop/DAT 400/UNRATE.xls")
str(UnemploymentRate) ## Checking the data types for my variables
## Classes 'tbl_df', 'tbl' and 'data.frame': 860 obs. of 2 variables:
## $ observation_date: POSIXct, format: "1948-01-01" "1948-02-01" ...
## $ UNRATE : num 3.4 3.8 4 3.9 3.5 3.6 3.6 3.9 3.8 3.7 ...
colnames(UnemploymentRate)[colnames(UnemploymentRate)== "observation_date"] <- "Date" ##Renaming a column for easier coding later on
UnemploymentRate$Date <- as.Date(UnemploymentRate$Date) ## Changed this variable to data type Date
UnemploymentRate$Year <- format(as.Date(UnemploymentRate$Date, "%d/%m/%Y"), "%Y")
UnemploymentRate$Year <- case_when(
UnemploymentRate$Year >= 1950 & UnemploymentRate$Year < 1960 ~ "1950-1960",
UnemploymentRate$Year >= 1960 & UnemploymentRate$Year < 1970 ~ "1960- 1970",
UnemploymentRate$Year >= 1960 & UnemploymentRate$Year < 1980 ~ "1970 -1980",
UnemploymentRate$Year >= 1980 & UnemploymentRate$Year < 1990 ~ "1980 - 1990",
UnemploymentRate$Year >= 1990 & UnemploymentRate$Year < 2000 ~ "1990-2000",
UnemploymentRate$Year >= 2000 & UnemploymentRate$Year < 2010 ~ "2000- 2010",
UnemploymentRate$Year >= "2010" ~ "2010 -2019"
)
NewGroupByTable <- sqldf("SELECT Year, avg(UNRATE) As AVG_RATE, Date
FROM UnemploymentRate
GROUP BY Year
"
)
na.omit(NewGroupByTable$Year)
## [1] "1950-1960" "1960- 1970" "1970 -1980" "1980 - 1990" "1990-2000"
## [6] "2000- 2010" "2010 -2019"
## attr(,"na.action")
## [1] 1
## attr(,"class")
## [1] "omit"
plotlab <- c("1950-1960", "1960-1970", "1970-1980", "1980-1990", "1990-2000","2000-2010", "2010-2019")
ggplot(data = NewGroupByTable, aes(x = Date, y = AVG_RATE ))+
########This section of code adds the recession shading, has to be before the plot so it does not cover the plotted lines #################################
geom_rect(aes(xmin = ymd('1948-11-01'),
xmax = ymd('1949-10-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1953-07-01'),
xmax = ymd('1954-05-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1957-08-01'),
xmax = ymd('1958-04-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1960-04-01'),
xmax = ymd('1961-02-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1969-12-01'),
xmax = ymd('1970-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1973-11-01'),
xmax = ymd('1975-03-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1980-02-01'),
xmax = ymd('1980-07-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1981-07-01'),
xmax = ymd('1982-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('1990-07-01'),
xmax = ymd('1991-03-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('2001-03-01'),
xmax = ymd('2001-11-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_rect(aes(xmin = ymd('2008-12-01'),
xmax = ymd('2009-06-01'),
ymin= -Inf,
ymax = Inf), fill = 'grey', alpha = .05)+
geom_line(color = "blue") +
geom_line(color = "blue") +
########This Section of code changes the tick start date on the x & y axis ##################################################
scale_x_date( date_labels = ( "%Y"), breaks= as.Date(c("1950-01-01", "1960-01-01","1970-01-01","1980-01-01","1990-01-01","2000-01-01","2010-01-01")),
labels= c(plotlab)) +
scale_y_continuous(breaks = (1:11)) +
##### This section of code adds the proper titles #####################################
labs(
x = "Date",
y = "Percent",
title = "Civilian Unemployment Rate",
subtitle = "Source: https://fred.stlouisfed.org"
)
library(readxl)
library(ggplot2)
library(lubridate)
library(dplyr)
library(plyr)
library(sqldf)
UnemploymentRate <- read_excel("C:/Users/shahi/Desktop/DAT 400/UNRATE.xls")
str(UnemploymentRate) ## Checking the data types for my variables
## Classes 'tbl_df', 'tbl' and 'data.frame': 860 obs. of 2 variables:
## $ observation_date: POSIXct, format: "1948-01-01" "1948-02-01" ...
## $ UNRATE : num 3.4 3.8 4 3.9 3.5 3.6 3.6 3.9 3.8 3.7 ...
colnames(UnemploymentRate)[colnames(UnemploymentRate)== "observation_date"] <- "Date" ##Renaming a column for easier coding later on
UnemploymentRate$Date <- as.Date(UnemploymentRate$Date) ## Changed this variable to data type Date
UnemploymentRate$Year <- format(as.Date(UnemploymentRate$Date, "%d/%m/%Y"), "%Y")
UnemploymentRate$Year <- case_when(
UnemploymentRate$Year >= 1950 & UnemploymentRate$Year < 1960 ~ "1950s",
UnemploymentRate$Year >= 1960 & UnemploymentRate$Year < 1970 ~ "1960s",
UnemploymentRate$Year >= 1960 & UnemploymentRate$Year < 1980 ~ "1970s",
UnemploymentRate$Year >= 1980 & UnemploymentRate$Year < 1990 ~ "1980s",
UnemploymentRate$Year >= 1990 & UnemploymentRate$Year < 2000 ~ "1990s",
UnemploymentRate$Year >= 2000 & UnemploymentRate$Year < 2010 ~ "2000s",
UnemploymentRate$Year >= "2010" ~ "2010s"
)
NewGroupByTable <- sqldf("SELECT Year, avg(UNRATE) As AVG_RATE, Date
FROM UnemploymentRate
GROUP BY Year
"
)
NewGroupByTable$Greater<- case_when(
NewGroupByTable$AVG_RATE > 5 ~"Yes",
TRUE ~ "No"
)
na.omit(NewGroupByTable)
## Year AVG_RATE Date Greater
## 2 1950s 4.511667 1950-01-01 No
## 3 1960s 4.779167 1960-01-01 No
## 4 1970s 6.217500 1970-01-01 Yes
## 5 1980s 7.272500 1980-01-01 Yes
## 6 1990s 5.762500 1990-01-01 Yes
## 7 2000s 5.541667 2000-01-01 Yes
## 8 2010s 6.312069 2010-01-01 Yes
ggplot(na.omit(NewGroupByTable), aes(x = Year , y = AVG_RATE, fill = Greater) )+
geom_bar(stat= "identity", width = .7)+
theme_minimal()+
labs(fill = "Greater Than 5%")+
labs(
x = "Decades",
y ="Unemployment Averages",
title = "America's Average Unemployment Per Decade",
subtitle = "Source: https://fred.stlouisfed.org"
)
