Currently when we want to count entries in a semantic domain or check whether a semantic domain has any, we have to query the Frontier database, checking through the semantic domains of the senses of the entries. For performant project statistics, we could create a new collection that keeps track of number of senses for each semantic domain. Each database entry would have projectId, domainId, and senseCount. This count would be updated any time a word is added, modified, or deleted.