# Use official BLS annual unemployment data vs manually calculating the average
# Source:
read_excel("~/Data/annual.xlsx", skip=10) %>%
  mutate(Year=as.character(as.integer(Year)), Annual=Annual/100) -> annual_rate

# The data source Andy Kriebel curated for you/us:!AhZVJtXF2-tD1UVEK7gYn2vN5Hxn #ty Andy!
read_excel("~/Data/staadata.xlsx") %>%
  left_join(annual_rate) %>%
  filter(State != "District of Columbia") %>%
    year = as.Date(sprintf("%s-01-01", Year)),
    pct = (Unemployed / `Civilian Labor Force Population`),
    us_diff = -(Annual-pct),
    col = ifelse(us_diff<0,
               "Better than U.S. National Average",
               "Worse than U.S. National Average")
  ) -> df

credits <- "Notes: Excludes the District of Columbia. 2016 figure represents October rate.\nData: U.S. Bureau of Labor Statistics <>\nCredit: Matt Stiles/The Daily Viz <>"
ggplot(df, aes(year, us_diff, group=State)) +
  geom_segment(aes(xend=year, yend=0, color=col), size=0.5) +
  scale_x_date(expand=c(0,0), date_labels="'%y") +
  scale_y_continuous(expand=c(0,0), label=scales::percent, limit=c(-0.09, 0.09)) +
  scale_color_manual(name=NULL, expand=c(0,0),
                     values=c(`Better than U.S. National Average`="#4575b4",
                              `Worse than U.S. National Average`="#d73027")) +
  facet_wrap(~State, ncol=5, scales="free_x") +
  labs(x=NULL, y=NULL, title="The State of U.S. Jobs: 1976-2016",
       subtitle="Percentage points below or above the national unemployment rate, by state. Negative values represent unemployment rates\nthat were lower — or better, from a jobs perspective — than the national rate.",
       caption=credits) +
  theme_hrbrmstr_msc(grid="Y", strip_text_size=9) +
  theme(panel.background=element_rect(color="#00000000", fill="#f0f0f055")) +
  theme(panel.spacing=unit(0.5, "lines")) +
  theme(plot.subtitle=element_text(family="MuseoSansCond-300")) +