How to create a chloropleth map in R Plotly based on a Categorical variable?

Question

I'm trying to create a chloropleth map of the US that uses a categorical variable for the state color, but I only get a blank map. Do plotly maps have compatibility with categorical data? If so, how does the syntax change?

For my data, I'm simply uploading a table of rows consisting the state and randomly one of "Good", "Bad", "OK."

What can I change in the code below for it to work? I've tried a workaround that slightly works to change the states' color but the colorbar gets wonky. (value4 is my Categorical Variable of "Good", "Bad", "OK")

Apologies if my question is not clear or my info is not great. I can answer further questions if anyone has them. Thanks in advance

foo <- brewer.pal(n = 3,
                        name = "Set1")

df <- mutate(df, test = ntile(x = value4, n = 3))

cw_map <- plot_ly(
  data = df,
  type = "choropleth",
  locations = ~ state,
  locationmode = "USA-states",
  color = ~ test,
  colors = foo[df$test],
  z = ~ test
) %>%
  layout(geo = list(scope = "usa"))

print(cw_map)

Answer 1

You need to have the states in the code form so let's start with that:

STATES <-c("AL", "AK", "AZ", "AR", "CA", "CO", "CT", "DE", "FL", "GA", 
"HI", "ID", "IL", "IN", "IA", "KS", "KY", "LA", "ME", "MD", "MA", 
"MI", "MN", "MS", "MO", "MT", "NE", "NV", "NH", "NJ", "NM", "NY", 
"NC", "ND", "OH", "OK", "OR", "PA", "RI", "SC", "SD", "TN", "TX", 
"UT", "VT", "VA", "WA", "WV", "WI", "WY")

Like you did, we give random value4 for each state:

df = data.frame(state=STATES,
value4=sample(c("Good", "Bad", "OK."),length(STATES),replace=TRUE))

Then we make your value4 as factor, and colors etc as you have done before:

df$value4 = factor(df$value4)
df$test = as.numeric(df$value4)
nfactor = length(levels(df$value4))
foo <- brewer.pal(n = nfactor,name = "Set1")
names(foo) = levels(df$value4)

To have the color legend in discrete form, you need to provide it as a data frame that defines your breaks on the relative scale for z. It's not very well documented in R plotly and I wrote the solution below for n factors with information from @emphet's plotly forum post and @marcosandri's SO post:

Z_Breaks = function(n){
CUTS = seq(0,1,length.out=n+1)
rep(CUTS,ifelse(CUTS %in% 0:1,1,2))
}

colorScale <- data.frame(z=Z_Breaks(nfactor),
col=rep(foo,each=2),stringsAsFactors=FALSE)

          z     col
1 0.0000000 #E41A1C
2 0.3333333 #E41A1C
3 0.3333333 #377EB8
4 0.6666667 #377EB8
5 0.6666667 #4DAF4A
6 1.0000000 #4DAF4A

And we plot:

cw_map <- plot_ly(
  data = df,
  type = "choropleth",
  locations = ~ state,
  locationmode = "USA-states",
  z = df$test,
  colorscale=colorScale,
  colorbar=list(tickvals=1:nfactor, ticktext=names(foo))
) %>%
layout(geo = list(scope = "usa"))