Question

I am interested in displaying “large n” ordinal scale survey response information across 2 covariates (consumer attribute and provider attribute). I am using R’s hexbin package. I would like to have the size of the hexagon indicate the number of survey responses and the color of the hexagon to represent the linear mean satisfaction response.

I can’t seem to find documentation indicating how one would do so with either the hexbin or other existing packages.

EDIT for more clarification:

My data is structured like the following:

csr_score pro_score rating
       15        16      8
       17        18     10
       19        12      4

where the length of this data frame is 20,000 rows.

Answer 1

The key function you’re looking for is hexTapply:

df <- data.frame (c = runif (1000), p = runif (1000), rating = rnorm (1000))
h <- hexbin (x=df$c, y = df$p, IDs = TRUE, xbins=5)
rating.binned <- hexTapply (h, df$rating, FUN=mean)

df.binned <- data.frame (c = h@xcm, p = h@ycm, freq = h@count, rating = rating.binned)
ggplot (df.binned, aes (x = c, y = p, col = rating, size = freq)) + geom_point ()