Separately, it would be cool to have a read-only API for seeing what other objects (and therefore selected data) are on the page. Use case: our extension could list existing objects on the page and have the user select which one to use as a data source, instead of them configuring our visual with the same dimensions/measures.