Unlocking the power of data visualization with R

Gazzelloni F. (2023), Data Visualization: Lyme Disease Map

Load libraries

library(tidyverse)

Data for this visualization is from CDC: https://www.cdc.gov/lyme/stats/survfaq.html

This is old dataset, it is updated at 2011. I had a look at it, then used the most updated one.

library(jsonlite)

url <- "https://data.cdc.gov/api/views/smai-7mz9/rows.json?accessType=DOWNLOAD"
data <- fromJSON(url)
keys <- data$meta$view$columns

df <- as.data.frame(data$data)
names(df) <- keys$name%>%t()
df%>%head

my_df <- df %>%
  select(StateCode,CountyCode,StateName,CountyName,
         ConfirmedCount_1992_1996,
         ConfirmedCount_1997_2001,
         ConfirmedCount_2002_2006,
         ConfirmedCount_2007_2011)%>%
  pivot_longer(5:8)%>%
  mutate(name=gsub("ConfirmedCount_","",name),
         name=gsub("_","-",name),
         value=as.numeric(value),
         value=ifelse(is.na(value),0,value),
         CountyName=tolower(CountyName),
         CountyName=gsub(" county","",CountyName),
         CountyName=trimws(CountyName))%>%
  rename(Years=name,subregion=CountyName)


my_df%>%head

This data contains County-level Lyme disease data from 2001-2020

url2 <- "https://www.cdc.gov/lyme/resources/datasurveillance/LD-Case-Counts-by-County-01-20.csv"

df2 <- read.csv(url2)

Tidy data, the Ctyname vector is the one I want to use for looking at the cases by cities. I had to leave some information behind.

my_df2 <- df2 %>%
  select(1:2,6:25)%>%
  pivot_longer(3:22)%>% 
  mutate(name=gsub("Cases","",name),
         Stname=tolower(Stname),
         Ctyname= str_to_lower(Ctyname),
         Ctyname=gsub(" county","",Ctyname),
         Ctyname=gsub(" parish","",Ctyname),
         Ctyname=gsub(" city","",Ctyname),
         Ctyname=gsub("st.","st",Ctyname),
         Ctyname=gsub(" bay borough","",Ctyname),
         Ctyname=case_when(Ctyname=="armstong"~"armstrong",
                           TRUE~Ctyname),
         name=as.numeric(name),
         name=cut_interval(name,4))%>%
  rename(Years=name,subregion2=Ctyname,region=Stname)%>%
  distinct()

Here is a double check of the City names differences in map_data("county") and the Lyme disease dataset (my_df2). I am going to leave some city names behind.

us_county_map <- map_data("county")
subregion <- us_county_map%>%count(subregion)%>%select(-n)%>%unlist()
subregion2 <- my_df2%>%count(subregion2)%>%select(-n)%>%unlist()
intersect(subregion, subregion2)%>%length()
setdiff(subregion2, subregion)%>%length()
setdiff(subregion, subregion2)%>%length()

my_df2%>%
  filter(str_detect(subregion2,"yellowstone"))

These are the sets that will be used in the Map.

Base map layers

us_state_map <- map_data("state")
us_county_map <- map_data("county")

Cases layer

my_us_county_map <- map_data("county")%>%
  filter(subregion%in%subregion2)

my_df_coords <- my_df2%>%
  rename(subregion=subregion2)%>%
  inner_join(my_us_county_map,by=c("region","subregion"))%>%
  distinct()

df_jitter <- my_df_coords%>%
  filter(value>0)%>%
  group_by(subregion,region,Years)%>%
  mutate(long=mean(range(long)),lat=mean(range(lat)))%>%
         #value=sum(value))%>%
  ungroup()%>%
  distinct()

Change the facet lables:

lbl <- my_df2 %>% 
  group_by(Years)%>%
  reframe(value=sum(value))%>%
  mutate(pct=lag(value),
         pct=round((value-pct)/pct*100,2),
         pct=ifelse(is.na(pct),"-",pct),
         Year=gsub("\\[|\\]", "", Years),
         Year=gsub("\\(|\\]", "", Year),
         Year=gsub(",", "-", Year),
         lbl=paste(Year,"pct change",pct,"%"))%>%
  select(lbl)%>%
  unlist()

Set the fonts:

library(showtext)
library(sysfonts)
sysfonts::font_add_google("David Libre","David Libre")
sysfonts::font_add_google("Syne Mono","Syne Mono")

Make the Map:

showtext_auto()
p <-ggplot()+
  geom_polygon(data=us_county_map,
             aes(x=long,y=lat,group = group),
             fill="white",color = "#192d40",linewidth=0.02) +
  geom_polygon(data=us_state_map,aes(x=long,y=lat,group = group),
             fill=NA,color = "grey50",linewidth=0.2)+
  geom_jitter(data=df_jitter,
               aes(x=long,y=lat,size=value),
              color = "#192d40")+
  coord_quickmap()+
  facet_wrap(~Years,scales = "free",
             labeller = as_labeller(setNames(lbl, sort(unique(df_jitter$Years))))) +
  guides(size="none")+
  scale_size(range=c(0.1,1))+  
  labs(caption="#30DayChartChallenge Day 29 Monochrome\nDataSource: CDC Lyme Disease | Map: Federica Gazzelloni") +
  ggthemes::theme_map()+
  theme(text=element_text(color="white",size=40),
        legend.background = element_blank(),
        legend.position = "right",
        strip.background = element_rect(color=NA,fill=NA),
        strip.text = element_text(family="David Libre",color="white"),
        plot.background = element_rect(color="#192d40",fill="#192d40"),
        plot.caption = element_text(family="Syne Mono",lineheight = 0.5,hjust=0.3),
        panel.background = element_rect(color="#192d40",fill="#192d40"),
        plot.margin = margin(0,0,0,0,unit = "pt"))

Have a look at the Map:

Set some information about the Lyme disease, and use ggtitle() for making a title plot to use in the main plot layout.

library(ggtext)
text <- "Lyme disease is caused by bacteria called Borrelia burgdorferi, which is spread to people through the bite of several types of blacklegged ticks. In the United States, most infections occur in Eastern states, Northern midwestern states, and West Coast. Recent estimates using other methods suggest that approximately 476,000 people may get Lyme disease each year in the United States."

text=tibble(text)

title <- ggplot() + 
ggtitle(label =text) +
  theme(
    plot.title.position = "plot",
    plot.title = element_textbox_simple(
      size = 50,lineheight = 0.3,
      color="white",
      family="Syne Mono",
      padding = margin(5.5, 5.5, 5.5, 5.5),
      margin = margin(0, 0, 5.5, 0),
      fill = "#192d40"),
    plot.background = element_rect(color="#192d40",fill="#192d40"),
        panel.background = element_rect(color="#192d40",fill="#192d40"),
    plot.margin = margin(0,0,0,0,unit = "pt")
    )

Here is the part where the layout design is done. Use {patchwork} for setting the area to combine the title with the map with the area() function. Then check the result with the plot() function, as shown below:

library(patchwork)
# area(t, l, b = t, r = l)
design <- c(
   area(1, 1,20,100),
   area(5, 1, 100,100)
)

plot(design)

Finally, combine all together in one plot:

title + p + plot_layout(design = design) &
  theme(plot.background = element_rect(fill="#192d40",color="#192d40"))

And save it!

ggsave("p.png",width = 9,height = 7,dpi=320,bg="#192d40")