new_brief_figures_EPS.Rmd

---
title: "carsey"
author: "Coda Rayo-Garza"
date: "3/23/2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
require("knitr")
opts_knit$set(root.dir = "~/Stats II/Data")
library(ipumsr,quietly = T)
library(dplyr,quietly = T)
library(car,quietly = T)
library(zoo,quietly = T)
library(ggplot2,quietly = T)
library(questionr,quietly = T)
library(survey,quietly = T)

rm(list=ls()); gc()
```

# Pct change data load
```{r}
sasy<-haven::read_sas("~/OneDrive - University of Texas at San Antonio/projects/carsey_brief/cps_00040.sas7bdat.gz")

load("C:/Users/ozd504/OneDrive - University of Texas at San Antonio/projects/carsey_brief//cpsbriefdata.Rdata")
#View(cpsdat)
source("C:/Users/ozd504/Documents/GitHub/custom_colors//carsey_colors.R")
names(cpsdat) <- tolower(names(cpsdat))
#names(cpsdat)

##describe(cpsdat$region)

#Create a weighted freq table of lf part by month by region 
cpsdat$yr_month <- as.Date(as.yearmon(paste(cpsdat$year,cpsdat$month, sep="/"), format = "%Y/%m"))
cpsdat$reg<- car::recode(cpsdat$region, recodes =" 11:12='Northeast' ; 21:22='North Central' ; 31:33='South'; 41:42='West';97=NA", as.factor=T)
cpsdat$educ2<-car::Recode(cpsdat$educ, recodes = "21:110='< College';  111:125='Bachelors +';else=NA; 0=NA", as.factor=T)

cpsdat<- cpsdat%>%
  mutate(emp = Recode(empstat, recodes = "0 = NA; 1='af'; 10='curr work'; 12 = 'recent no job'; 20:22='unemp'"),
         sex=Recode(sex, recodes = "1='Male'; 2='Female'; else=NA", as.factor=T),
         race = Recode (race, recodes = "100 = 'White'; 200='Black';300='AIAE';651='Asian';652='NHPI';802:830='multiple'; else=NA"),
         hisp= ifelse(hispan !=0, "Latino", "Not Latino"),
         nativity = ifelse(citizen%in%c(1:3), 1, 0),
        curremp = ifelse(empstat%in%c(10,12) , 1, 0),
          recentloss=ifelse(emp=='recent no job', 1,0))

cpsdat$emp_bin<-ifelse(cpsdat$empstat%in%c(10,11, 12), 1, 0)
cpsdat$race_eth<-interaction(cpsdat$race, cpsdat$hisp)
library(stringr)
cpsdat$race_eth2<-ifelse(str_sub(as.character(cpsdat$race_eth), start = -10)=="Not Latino", as.character(cpsdat$race_eth),"Latino")

# cpsdat<-cpsdat%>%
#   filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", race_eth2!="Asian.Not Latino", age>=16)

des<-survey::svydesign(ids=~1,  weights = ~wtfinl , data=cpsdat[is.na(cpsdat$wtfinl)==F&cpsdat$wtfinl>0,])

#totm<-data.frame(svytable(~yr_month+emp_bin, des))
#write.csv(cpsdat[cpsdat$race_eth2=="Latino",], "~/cpsdat.csv", na = ".")
```

```{r,eval=FALSE}
test<- cpsdat%>%
  filter(emp_bin==1,  race_eth2=="Latino", is.na(educ2)==F )%>%
  group_by(race_eth2, sex, nativity, yr_month, educ2)%>%
  summarize(ns = n())
  
table(test$ns >=30)


```

# figure 1
```{r}
total<-data.frame(svytable(~yr_month+emp_bin,des))


totalyr_month<- as.Date(total$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- total %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select( february)

lfpartreg$february<-febreg$february


#str(lfpartreg)

#library(forcats)
feb <- lfpartreg$Freq[lfpartreg$yr_month=="2020-02-01"]
may <- lfpartreg$Freq[lfpartreg$yr_month=="2021-06-01"]
diff<- lfpartreg$Freq[lfpartreg$yr_month=="2020-02-01"] - lfpartreg$Freq[lfpartreg$yr_month=="2021-06-01"]

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq))%>%
  ggplot(aes(x=yr_month, y=pct_ch, group="Change"))+
  geom_line(color=carsey_colors[1], lwd=1.15)+
  geom_hline(yintercept = max(lfpartreg$Freq), lwd=.7, lty=2)+
  geom_hline(yintercept = lfpartreg$Freq[lfpartreg$yr_month=="2021-06-01"], lwd=.7, lty=2)+
  geom_segment(aes(x= "2021-06-01", y= feb, xend = "2021-06-01", yend = may ),
               arrow = arrow(ends = "both", length = unit(.2, "cm")))+
  # geom_point()
  #  geom_line(aes(group = yr_month),
  #           col = "grey")+
  annotate("text", x = "2021-05-01", y = 155000000, label = "-5,851,544")+
  annotate("text", x = "2021-01-01", y = 157402659, label = "Pre Pandemic")+
annotate("text", x = "2021-01-01", y = 154000000, label = "June 2021")+
  #facet_wrap(~sex)+
  labs(title = "Number of Employed People per Month",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format(scale = .000001))
  
 ggsave("images/fig1.eps",height=8, width=10, dpi = "print")

```

# figure 2
```{r}
byreg <- data.frame(svytable(~yr_month+emp_bin+sex,des))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(sex, february)

lfpartreg<- left_join(lfpartreg, febreg, by="sex")


#str(lfpartreg)

library(forcats)


pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = sex, color=factor(sex)), lwd=1.15)+
  geom_hline(yintercept = 0, lwd=.7,lty=2)+
  labs(title = "Percent Change in Number of Employed Persons from\nFebruary 2020 by Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
   scale_color_manual( name = "Gender",labels=c("Female", "Male"), values = c("#004f9e","#e36c2c","#4eaf49"))

 ggsave("images/fig2.eps",height=8, width=10, dpi = "print")
```

# figure 3
```{r}
byreg <- data.frame(svytable(~yr_month+emp_bin+race_eth2,des))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01'),race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino" )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(race_eth2, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c( "race_eth2"))


#str(lfpartreg)

#library(forcats)
pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)


lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = race_eth2, color=factor(race_eth2)), lwd=1.15)+
  geom_hline(yintercept = 0, lwd=.7,lty=2)+
  labs(title = "Percent Change in Number of Employed Persons from\nFebruary 2020 by Race/Ethnicity",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data\n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  #facet_wrap(~sex)+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  #scale_color_carsey(name = "Race/Ethnicity", labels=c("Black", "Latino", "White"))
 scale_color_manual( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"), values = c("#004f9e","#e36c2c","#4eaf49"))
 ggsave("images/fig3.eps",height=8, width=10, dpi = "print")
```



# figure 5
```{r}
byreg <- data.frame(svytable(~yr_month+emp_bin+sex+race_eth2,des))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01'), race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino" )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(sex, race_eth2, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c("sex",  "race_eth2"))


pcts<-lfpartreg%>%
  mutate(pct_ch = (Freq/february)*100-100)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = race_eth2, color=factor(race_eth2)), lwd=1.15)+
 # geom_hline(yintercept = 0, lwd=.7,lty=2)+
  labs(title = "Percent Change in Number of Employed Persons from\nFebruary 2020 by Race/Ethnicity and Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  facet_wrap(~sex)+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
   scale_color_manual( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"), values = c("#004f9e","#e36c2c","#4eaf49"))

 ggsave("images/fig5.eps",height=8, width=10, dpi = "print")
```


# figure 6
```{r}
des2<-subset(des, race_eth2 =="Latino")

byreg <- data.frame(svytable(~yr_month+emp_bin+nativity+sex,des2))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(nativity,sex, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c("nativity", "sex"))


pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)


library(forcats)

out<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = nativity, color=factor(nativity)), lwd=1.15)+
  #geom_hline(yintercept = 0, lwd=.7,lty=2)+
  facet_wrap(~sex)+
  labs(title = "Percent Change in Number of Latina/o Employed Persons from\nFebruary 2020  by Nativity and Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data\n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  scale_color_carsey(name = "Nativity", labels =c("Foreign Born", "Native Born"))

 ggsave("images/fig6.eps",height=8, width=10, dpi = "print")
```

# figure 9 pct change by gender and education
```{r,eval=FALSE}
byrace<-data.frame(svytable(~yr_month+emp_bin+educ2,des))


byrace$yr_month<- as.Date(byrace$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byrace %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select( educ2, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c( "educ2"))

pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)

#str(lfpartreg)

#library(forcats)

lfpartreg%>%
   
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = educ2, color=educ2), lwd=1.15)+
  geom_hline(yintercept = 0, lty=2, lwd=.7)+
  #facet_wrap(~sex)+
  labs(title = "Percent Change in Number of Latina/o Employed Persons from\nFebruary 2020 by Education Level",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  #scale_y_continuous(labels = scales::number_format())+
  #scale_color_carsey(name = "Education", labels= c("< College", "Bachelors +"))
  scale_color_manual( name = "Education",
                      labels= c("< College", "Bachelors +"), values = c("#004f9e","#e36c2c"))

 ggsave("images/fig9.eps",height=8, width=10, dpi = "print")

```



# figure 10 pct change by gender, educ and race/eth
```{r fig.height=8, fig.width=10}
byrace<-data.frame(svytable(~yr_month+emp_bin+educ2+race_eth2+sex,des))


byrace$yr_month<- as.Date(byrace$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byrace %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select( educ2,race_eth2,sex,february)

lfpartreg<- left_join(lfpartreg, febreg, by=c("sex","educ2", "race_eth2"))


pcts<-lfpartreg%>%
  mutate(pct_ch = (Freq/february)*100-100)

lfpartreg%>%
  filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino")%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = race_eth2, color=race_eth2), lwd=1.15)+
  geom_hline(yintercept = 0, lty=2, lwd=.7)+
  facet_wrap(~educ2+sex)+
  labs(title = "Percent Change in Number of Employed Persons from\nFebruary 2020 by Race/Ethnicity, Gender and Education Level February 2020 to 2021
",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  #scale_color_carsey(name = "Education", labels= c("< College", "Bachelors +"))
   scale_color_manual( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"), values = c("#004f9e","#e36c2c","#4eaf49"))

 ggsave("images/fig10.eps",height=8, width=10, dpi = "print")

```

# figure 12 pct change among latinos by educ,nativity and gender

```{r, fig.height=8, fig.width=10}
des2<-subset(des, race_eth2 =="Latino")



byreg <- data.frame(svytable(~yr_month+emp_bin+nativity+sex+educ2,des2))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(nativity,sex,educ2, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c("nativity","educ2", "sex"))


pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)


library(forcats)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = nativity, color=factor(nativity)), lwd=1.15)+
  geom_hline(yintercept = 0, lwd=.7,lty=2)+
  facet_wrap(~sex+educ2)+
  labs(title = "Change in Number of Employed People by Nativity and Gender among Latina/os",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data\n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Change in Number of Employed Persons")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  scale_color_carsey(name = "Nativity", labels =c("Foreign Born", "Native Born"))

 ggsave("images/fig12.eps",height=8, width=10, dpi = "print")
```



## unemployment rate figures
# data prep
# Load data
```{r}
cpsdat<-readRDS("data/cpsdat_newbrief.rds")
```

## recoding from original data
```{r,  echo=FALSE}
cpsdat<-cpsdat%>%
  filter(YEAR >2009,AGE>16,EMPSTAT%in%c(10,12,21,22))%>%
  mutate(emp = Recode(EMPSTAT, recodes = "0 = NA; 1='af'; 10='curr work'; 12 = 'recent no job'; 20:22='unemp'"),
         sex=Recode(SEX, recodes = "1='Male'; 2='Female'; else=NA", as.factor=T),
         race = Recode (RACE, recodes = "100 = 'White'; 200='Black';300='AIAE';651='Asian';652='NHPI';802:830='multiple'; else=NA"),
         lfpart = Recode(LABFORCE,recodes="2=1; 1=0; else=NA" ),
         hisp= ifelse(HISPAN !=0, "Latino", "Not Latino")
         )%>%
  mutate( curremp = ifelse(EMPSTAT%in%c(10,12) , 1, 0),
          recentloss=ifelse(emp=='recent no job', 1,0))%>%
  arrange(CPSIDP, MONTH,MISH)

cpsdat$emp_bin<-ifelse(cpsdat$emp!="curr work", 1, 0)
cpsdat$month<- as.Date(as.yearmon(paste(cpsdat$YEAR,cpsdat$MONTH, sep="/"), format = "%Y/%m"))
cpsdat$race_eth<-interaction(cpsdat$race, cpsdat$hisp)
library(stringr)
cpsdat$race_eth2<-ifelse(str_sub(as.character(cpsdat$race_eth), start = -10)=="Not Latino", as.character(cpsdat$race_eth),"Latino")
#cpsdat$fb<-ifelse(cpsdat$CITIZEN%in%c(4,5),"Non-citizen", "Citizen")
#cpsdat
cpsdat$byr<-2020-cpsdat$AGE
cpsdat$cohort<-car::Recode(cpsdat$byr, recodes = "1944:1964='boomer'; 1965:1979='genx'; 1980:1994='geny'; 1995:2015='genz'; else=NA")
cpsdat$educ<-car::Recode(cpsdat$EDUC, recodes = "21:72='< High School'; 73='High School'; 81:110='Some college';  111:125='Bachelors +';else=NA; 0=NA", as.factor=T)
cpsdat$duremp_c<-ifelse(cpsdat$DURUNEMP==999, NA, cpsdat$DURUNEMP)
cpsdat$fb<-ifelse(cpsdat$YRIMMIG>0, "Foreign Born", "Native Born")
cpsdat$avghrs<-ifelse(cpsdat$AHRSWORKT==999, NA, cpsdat$AHRSWORKT)
cpsdat<-cpsdat%>%
  select(emp_bin,CPSIDP, MISH, MONTH, WTFINL, month, race_eth2, curremp,educ, sex, race, emp, YEAR, AGE, STATEFIP, PANLWT, EDUC, avghrs, ABSENT, WKSTAT, EMPSAME, MULTJOB, NUMJOB,recentloss,fb, cohort , METRO, duremp_c, lfpart)%>%
  filter(is.na(race_eth2)==F, month>=as.Date('2020-02-01'))%>%
  #filter(AGE>18)%>%
  group_by(CPSIDP)%>%
  arrange(CPSIDP, MONTH)
```

# figure 4 unemployed by race
```{r}
levs<-levels(as.factor(cpsdat$race_eth2))
outna<-cpsdat%>%
  filter(is.na(race_eth2)==F, month>=as.Date('2020-02-01'))%>%
  #mutate(Group = ifelse(STATEFIP=="Texas", "Texas","Rest of US"))%>%
  group_by(month, race_eth2 )%>%
  summarise(#unemprate =(wtd.mean(recentloss,weights = WTFINL, na.rm=T)),
         # avghrs=wtd.mean(AHRSWORKT,weights = WTFINL, na.rm=T),
          emprate =(1- wtd.mean(curremp,weights =WTFINL, na.rm=T)))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2, month )

head(outna)

outna%>%
 filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", emprate!=0, race_eth2!="Asian.Not Latino")%>%
  ggplot()+geom_line(aes(x=month, y=emprate,color=race_eth2, group=race_eth2), lwd=2)+
  #scale_fill_discrete(name=)+
  labs(title = "Percent Unemployed by Race/Ethnicity",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Unemployed")+ylim(c(0, .2))+
  scale_color_manual( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"), values = c("#004f9e","#e36c2c","#4eaf49"))+
  scale_y_continuous(labels = scales::percent)+
  #guides(fill=guide_legend(title="Race/Ethnicity"))+
  #geom_hline(yintercept = 0, col="red", lwd=1.1)+
theme_minimal()+theme(axis.text.x = element_text(angle = 45))


ggsave(filename = "images/fig4.eps",height=8, width=10, dpi = "print" )

```
# figure 7 unemploy by gender and race
```{r, fig.width=10, fig.height=8}
outna2<-cpsdat%>%
  filter(is.na(race_eth2)==F)%>%
  #mutate(Group = ifelse(STATEFIP=="Texas", "Texas","Rest of US"))%>%
  group_by(month, race_eth2, sex )%>%
  summarise(#unemprate =(wtd.mean(recentloss,weights = WTFINL, na.rm=T)),
          #avghrs=wtd.mean(AHRSWORKT,weights = WTFINL, na.rm=T),
          emprate =(1- wtd.mean(curremp,weights = WTFINL, na.rm=T)))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2,sex, month )

head(outna2)

outna2%>%
    filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", emprate!=0, race_eth2!="Asian.Not Latino")%>%

  ggplot()+geom_line(aes(x=month, y=emprate,color=race_eth2, group=race_eth2), lwd=2)+
  #scale_fill_discrete(name=)+
  labs(title = "Percent Unemployed by Race/Ethnicity and Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Unemployed")+ylim(c(0, .25))+
  scale_color_manual( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"), values = c("#004f9e","#e36c2c","#4eaf49"))+facet_wrap(~sex)+scale_y_continuous(labels = scales::percent)+
  #guides(fill=guide_legend(title="Race/Ethnicity"))+
  #geom_hline(yintercept = 0, col="red", lwd=1.1)+
theme_minimal()+theme(axis.text.x = element_text(angle = 45))

#p$labels$colour<-"Race/Ethnicity"
#p
#<<<<<<< HEAD
ggsave(filename = "images/fig7.eps",height=8, width=10, dpi = "print" )

```


# figure 8 unemploy by gender and nativity
```{r, fig.width=10, fig.height=8}
outna2<-cpsdat%>%
  filter(is.na(race_eth2)==F)%>%
  #mutate(Group = ifelse(STATEFIP=="Texas", "Texas","Rest of US"))%>%
  group_by(month, race_eth2, fb, sex )%>%
  summarise(#unemprate =(wtd.mean(recentloss,weights = WTFINL, na.rm=T)),
         # avghrs=wtd.mean(AHRSWORKT,weights = WTFINL, na.rm=T),
          emprate =(1- wtd.mean(curremp,weights = WTFINL, na.rm=T)))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2,fb, month )

head(outna2)

outna2%>%
  mutate(group=paste(race_eth2, fb, sep = "-"))%>%
  filter( emprate!=0, race_eth2=="Latino")%>%
  ggplot()+geom_line(aes(x=month, y=emprate,color=fb, group=fb), lwd=2)+
  #scale_fill_discrete(name=)+
  labs(title = "Percent Unemployed Among Latinos by Nativity Status and Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Unemployed")+
  scale_color_manual( name = "Nativity Status",labels=c("Foreign Born", "Native Born"),
                      values = c("#004f9e","#e36c2c","#4eaf49"))+
  facet_wrap(~sex)+
  scale_y_continuous(labels = scales::percent)+
  #guides(fill=guide_legend(title="Race/Ethnicity"))+
  #geom_hline(yintercept = 0, col="red", lwd=1.1)+
theme_minimal()+theme(axis.text.x = element_text(angle = 45))

ggsave(filename = "images/fig8.eps",height=8, width=10, dpi = "print" )

```

# figure 11 unemp by educ race/eth and gender
```{r, fig.width=10, fig.height=8}
outna2<-cpsdat%>%
  filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino")%>%
  mutate(educ2 = ifelse(educ %in% c("< High School", "High School", "Some college"), "Not College Graduate", "College Graduate"))%>%
  #mutate(Group = ifelse(STATEFIP=="Texas", "Texas","Rest of US"))%>%
  group_by(month,race_eth2,  educ2, sex )%>%
  summarise(unemprate =(wtd.mean(recentloss,weights = WTFINL, na.rm=T)),
          #avghrs=wtd.mean(AHRSWORKT,weights = WTFINL, na.rm=T),
          emprate =(1- wtd.mean(curremp,weights = WTFINL, na.rm=T)),
          nsamp=n())%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( sex,educ2,race_eth2, month )

#write.csv(outna2,file = "docs/unemp_to_rogelio.csv", row.names = F)
head(outna2)
library(forcats)
outna2%>%
  filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino",nsamp>30, is.na(educ2)==F)%>%
   mutate(educ=fct_relevel(educ2,"Not College Graduate","College Graduate" ))%>%
  ggplot()+
  geom_line(aes(x=month, y=emprate,color=race_eth2, group=race_eth2), lwd=2)+
 # geom_hline(aes( yintercept = max(emprate), group=educ), lty=2, cex=.5)+
#  geom_hline(yintercept = min(outna2$emprate[outna$month=="2021-05-01"]), lty=2, cex=.5)+
  #scale_fill_discrete(name=)+
  labs(title = "Percent Unemployed by Race/Ethnicity, Gender and Education",
       subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Unemployed")+
  ylim(c(0, .4))+
  
 scale_color_carsey( name = "Education",labels=c(" Black", "Latino"," White"))+
  facet_wrap(~sex+educ, ncol = 2)+
  scale_y_continuous(labels = scales::percent)+
  #guides(fill=guide_legend(title="Race/Ethnicity"))+
  #geom_hline(yintercept = 0, col="red", lwd=1.1)+
theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))

ggsave(filename = "images/fig11.eps",height=8, width=10, dpi = "print" )

```

# figure 13 unemp among latinos by nativity, gender and educ
```{r, fig.width=10, fig.height=8}
outna2<-cpsdat%>%
  filter(is.na(race_eth2)==F)%>%
  mutate(educ2 = ifelse(educ %in% c("< High School", "High School", "Some college"), "Not College Graduate", "College Graduate"))%>%
  #mutate(Group = ifelse(STATEFIP=="Texas", "Texas","Rest of US"))%>%
  group_by(month, race_eth2, fb,educ2, sex )%>%
  summarise(#unemprate =(wtd.mean(recentloss,weights = WTFINL, na.rm=T)),
         # avghrs=wtd.mean(AHRSWORKT,weights = WTFINL, na.rm=T),
          emprate =(1- wtd.mean(curremp,weights = WTFINL, na.rm=T)))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2,fb,educ2, month )

head(outna2)
library(forcats)
outna2%>%
  #mutate(group=paste(race_eth2, fb, sep = "-"))%>%
  filter( emprate!=0, race_eth2=="Latino", is.na(educ2)==F)%>%
  mutate(educ=fct_relevel(educ2,"Not College Graduate","College Graduate" ))%>%
  ggplot()+geom_line(aes(x=month, y=emprate,color=fb, group=fb), lwd=2)+
  #scale_fill_discrete(name=)+
  labs(title = "Percent Unemployed Among Latinos by Nativity Status and Education",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data \n Calculations by Corey S. Sparks, Ph.D.",
       x = "Month",
       y = "% Unemployed")+ylim(c(0, .25))+
  scale_color_carsey( name = "Nativity",labels=c("Foreign Born", "Native Born"))+
  facet_wrap(~sex+educ2)+scale_y_continuous(labels = scales::percent)+
  #guides(fill=guide_legend(title="Race/Ethnicity"))+
  #geom_hline(yintercept = 0, col="red", lwd=1.1)+
theme_minimal()+theme(axis.text.x = element_text(angle = 45))

ggsave(filename = "images/fig13.eps",height=8, width=10, dpi = "print" )

```

## WFH figures
# load data
```{r}
load("~/OneDrive - University of Texas at San Antonio/projects/carsey_brief/cpsbriefdata.Rdata")
source("C:/Users/ozd504/Documents/GitHub/custom_colors/carsey_colors.R")
```


# Recode variables
```{r }
cpsdat<-cpsdat%>%
  mutate(month=as.Date(as.yearmon(paste(cpsdat$YEAR,cpsdat$MONTH, sep="/"), format = "%Y/%m")))%>%
  filter(month>=as.Date('2020-02-01'),AGE>16,EMPSTAT%in%c(10,12,21,22))%>%
  mutate(emp = Recode(EMPSTAT, recodes = "0 = NA; 1='af'; 10='curr work'; 12 = 'recent no job'; 20:22='unemp'"),
         sex=Recode(SEX, recodes = "1='Male'; 2='Female'; else=NA", as.factor=T),
         race = Recode (RACE, recodes = "100 = 'White'; 200='Black';300='AIAE';651='Asian';652='NHPI';802:830='multiple'; else=NA"),
         lfpart = Recode(LABFORCE,recodes="2=1; 1=0; else=NA" ),
         hisp= ifelse(HISPAN !=0, "Latino", "Not Latino")
         #remote = Recode (COVIDTELEW, recodes = "2=1; 1=0; else=NA" )
  )%>%
  mutate( curremp = ifelse(EMPSTAT%in%c(10,12) , 1, 0),
          recentloss=ifelse(emp=='recent no job', 1,0))%>%
  arrange(CPSIDP, MONTH,MISH)

#Remote variable
cpsdat$remote=ifelse(cpsdat$COVIDTELEW==2,1, ifelse (cpsdat$COVIDTELEW==1,0, NA))

cpsdat$emp_bin<-ifelse(cpsdat$emp!="curr work", 1, 0)
#cpsdat$month<- as.Date(as.yearmon(paste(cpsdat$YEAR,cpsdat$MONTH, sep="/"), format = "%Y/%m"))
cpsdat$race_eth<-interaction(cpsdat$race, cpsdat$hisp)
cpsdat$race_eth2<-ifelse(str_sub(as.character(cpsdat$race_eth), start = -10)=="Not Latino", as.character(cpsdat$race_eth),"Latino")
cpsdat$byr<-2021-cpsdat$AGE
cpsdat$cohort<-car::Recode(cpsdat$byr, recodes = "1944:1964='boomer'; 1965:1979='genx'; 1980:1994='geny'; 1995:2015='genz'; else=NA")
cpsdat$educ<-car::Recode(cpsdat$EDUC, recodes = "21:72='< High School'; 73='High School'; 81:110='Some college';  111:125='Bachelors +';else=NA; 0=NA", as.factor=T)
cpsdat$duremp_c<-ifelse(cpsdat$DURUNEMP==999, NA, cpsdat$DURUNEMP)
cpsdat$fb<-ifelse(cpsdat$YRIMMIG>0, "Foreign Born", "Native Born")
cpsdat$avghrs<-ifelse(cpsdat$AHRSWORKT==999, NA, cpsdat$AHRSWORKT)
cpsdat<-cpsdat%>%
  select(remote, emp_bin,CPSIDP,COVIDTELEW, MISH, MONTH, WTFINL, month, race_eth2, curremp,educ, sex, race, emp, YEAR, AGE, PANLWT, EDUC, avghrs, ABSENT, WKSTAT, EMPSAME, MULTJOB, NUMJOB,recentloss,fb, cohort , METRO, duremp_c, lfpart)%>%
  filter(is.na(race_eth2)==F, month>=as.Date('2020-05-01'))%>%
  group_by(CPSIDP)%>%
  arrange(CPSIDP, MONTH)

#Survey design
des<-survey::svydesign(ids=~1,  weights = ~WTFINL, data=cpsdat)
```

# figure 14 wfh by race

```{r cars}
levs<-levels(as.factor(cpsdat$race_eth2))
out_remote<-cpsdat%>%
  filter(is.na(race_eth2)==F, month>=as.Date('2020-05-01'))%>%
  group_by(month, race_eth2 )%>%
  summarise(
    remoterate =wtd.mean(remote,weights = WTFINL, na.rm=T))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2, month )
head (out_remote)

#eps("~/OneDrive - University of Texas at San Antonio/projects/carsey_brief/Figure1.eps")
out_remote%>%
  filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", remoterate!=0, race_eth2!="Asian.Not Latino")%>%
  ggplot()+geom_line(aes(x=month, y=remoterate,color=race_eth2, group=race_eth2), lwd=2)+
  labs(title = "Remote Work by Race/Ethnicity",subtitle = "May 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data ",
       x = "Month",
       y = "Percent of working remotely")+
  #ylim(c(0, .2))+
  scale_color_carsey( name = "Race/Ethnicity",labels=c(" Black", "Latino"," White"))+
  scale_y_continuous(labels = scales::percent, limits = c(0, .4))+
  theme_minimal()+theme(axis.text.x = element_text(angle = 45)) 

ggsave("images/fig14.eps",height=8, width=10, dpi = "print")
#rm(out_remout) 
```

# figure 15 wfh by race/gender
```{r cars}
out_gender<-cpsdat%>%
  filter(is.na(race_eth2)==F)%>%
  group_by(month, race_eth2, sex )%>%
  summarise(
    remoterate =(wtd.mean(remote,weights = WTFINL, na.rm=T)))%>%
  arrange( month)%>%
  ungroup()%>%
  arrange( race_eth2,sex, month )
head(out_gender)


out_gender%>%
  filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", remoterate!=0, race_eth2!="Asian.Not Latino")%>%
  ggplot()+geom_line(aes(x=month, y=remoterate,color=race_eth2, group=race_eth2), lwd=2)+
  labs(title = "Remote Work by Race/Ethnicity and Gender",subtitle = "May 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "Percent working remotely")+
  scale_color_carsey( name = "Race/Ethnicity",labels=c("Black", "Latino"," White"))+facet_wrap(~sex)+scale_y_continuous(labels = scales::percent, limits = c(0, .6))+
  theme_minimal()+theme(axis.text.x = element_text(angle = 45))

ggsave("images/fig15.eps",height=8, width=10, dpi = "print")
```

# figure 16 wfh by race/educ
```{r cars}
out_educ<-cpsdat%>%
   filter(is.na(race_eth2)==F,!is.na(educ))%>%
   group_by(month, race_eth2, educ)%>%
   summarise(
             remoterate =(wtd.mean(remote,weights = WTFINL, na.rm=T)))%>%
   arrange( month)%>%
   ungroup()%>%
   arrange( race_eth2,educ, month )
 
 library(forcats)


 out_educ%>%
   filter(race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino", remoterate!=0, race_eth2!="Asian.Not Latino")%>%
   
   mutate(educ=fct_relevel(educ,"< High School","High School","Some college","Bachelors +" ))%>%
   ggplot()+geom_line(aes(x=month, y=remoterate,color=race_eth2, group=race_eth2), lwd=1.1)+
   labs(title = "Remote work by Race/Ethnicity and Education",subtitle = "May 2020 to June 2021",
        caption = "Source: IPUMS CPS Monthly Data ",
        x = "Month",
        y = "Percent working remotely")+
   scale_color_carsey( name = "Race/Ethnicity",labels=c("Black", "Latino"," White"))+facet_wrap(~educ, ncol = 4)+
   scale_y_continuous(labels = scales::percent, limits = c(0, .6))+
   theme_minimal()+theme(axis.text.x = element_text(angle = 45))
ggsave("images/fig16.eps",height=8, width=10, dpi = "print")
```

# additional figures not used










```{r,eval=FALSE}
byreg <- data.frame(svytable(~yr_month+emp_bin+sex+race_eth2,des))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01'),race_eth2!="NHPI.Not Latino", race_eth2!="AIAE.Not Latino",race_eth2!="multiple.Not Latino",  race_eth2!="Asian.Not Latino" )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(sex, race_eth2, february)

lfpartreg<- left_join(lfpartreg, febreg, by=c("sex", "race_eth2"))

pcts<-lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)

#str(lfpartreg)

#library(forcats)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = race_eth2, color=factor(race_eth2)), lwd=1.15)+
  geom_hline(yintercept = 0, lwd=.7,lty=2)+
  labs(title = "Job Gains and Losses by Race/Ethnicity and Gender",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Jobs Recovered")+
  facet_wrap(~sex)+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  scale_color_carsey(name = "Race/Ethnicity", labels=c("Black", "Latino", "White"))

 ggsave("images/rogelio_race_gender.eps",height=8, width=10, dpi = "print")
```

```{r, eval=FALSE}
byreg <- data.frame(svytable(~yr_month+emp_bin+reg,des))
byreg$yr_month<- as.Date(byreg$yr_month)
#View(byreg)

#one for lfpart yes only 
lfpartreg <- byreg %>%
  filter(emp_bin==1,  yr_month!=('2020-01-01') )
 # mutate(
  #  prevmonth=lag(Freq,1), fmonth = ifelse(yr_month== ('2020-02-01'), 1, 0)) %>%
  #mutate(differnce=Freq-prevmonth)
#View(lfpartreg)

febreg<-lfpartreg%>%
  filter(yr_month==('2020-02-01'))%>%
  rename(february= Freq)%>%
  select(reg, february)

lfpartreg<- left_join(lfpartreg, febreg, by="reg")


#str(lfpartreg)

library(forcats)

lfpartreg%>%
 # filter(is.na(differnce)==F,  yr_month!=('2020-01-01'))%>%
  mutate(pct_ch = (Freq/february)*100-100)%>%
  ggplot()+
  geom_line(aes(x=yr_month, y=pct_ch, group = reg, color=factor(reg)), lwd=1.15)+
  geom_hline(yintercept = 0, lwd=.7,lty=2)+
  labs(title = "Job Gains and Losses by Census Region",subtitle = "February 2020 to June 2021",
       caption = "Source: IPUMS CPS Monthly Data",
       x = "Month",
       y = "% Jobs Recovered")+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45))+
  scale_y_continuous(labels = scales::number_format())+
  scale_color_carsey(name = "Region")

 ggsave("images/regional_jobs.eps",height=8, width=10, dpi = "print")
```