We noticed in our Data Clubs work that in a data set with a lot of missing data, the dots corresponding to missing data were colored gray – but that was also a color chosen for one of the values for the categorical variables. Seems like the color chosen for missing data should be deleted from the list of possible colors for attribute values.
You probably know this, but you can change the color corresponding to the category that has been assigned gray using the ‘brush’ menu on the graph inspector. (But you can’t change color assigned to missing values.)