We used this datastructure a few months ago to track an approximation to the cardinality of the followers of a set of twitter users, which is handy to calculate the reach of a twitter post. More recently, we used it again to track an approximation to the number of unique images that a user generated in a time period (each image is identified by a different URL).
Very handy data structure, and glad to see it land in Redis!
Very handy data structure, and glad to see it land in Redis!