We have several survey-based feature services where we'd like to filter the parent features, pass that filter to related tables of other data, and then provide charts of averages for the related data. If no data was collected for a given table (e.g. presence only survey), the average will be off, because the values are divided by the number of records in the related table, not the number of parent features.
Please add a modal menu for widgets that include the "AVERAGE" stat function to specify what count the sum of values will be divided by: the count of the current data source's visible rows, or the count of a different layer's visible rows. There are probably some other useful choices to include in a menu like this, but this is the distinction that's hanging us up right now.
