I'm seeing it as simply being open about internal workings mainly. I assume it is an internal database used as part of managing the project and Debian has always generally been a project that abhors keeping things under wraps.

For someone wanting to be more deeply involved than simply having Debian installed, access to a copy of this database could be a valuable learning and/or reference resource.

More abstractly, for someone learning/revising relational database concepts it could be a useful structure & dataset to practice on, particularly around the package dependency graph (a lot of example datasets out there as learning resources are hierarchical or just plain flat, so this is a welcome addition to that arena), and if they are at all interested in Debian then it is an example dataset/structure that is more relevant to them than a lot of other examples would be.

The wiki page lists some of the use-cases.