When a new report (prebuilt or custom) is created for development purposes, it is often not very useful unless you connect to an actual data source that produces real data.

While the ERD is very useful to understand the overall schema, tables, columns etc, it would be even better to have a "Sandbox" mode for connectors, which would populate the destination tables with realistic data. That data can then be build to build & test the next stage of your ELT pipelines, before creating a connector that ingests "real" data.