Right now WikiProject article assessment metadata is all stored in WikiText in templates on article talk pages (although some of it can also be accessed via categories). This means that there are no easy ways to query this data and it has to be aggregated and reported through scripts and bots.
A better solution would be to have a parser function that could be embedded in WikiProject templates and would map common template parameters like class and importance to corresponding columns in a database table. So for example, the template WikiProject Medicine on English Wikipedia could include a call like so in the template code:
{{wikiproject:Medicine|{{{class}}}|{{{importance}}}}}
Then each time the template was transcluded on an article talk page, it would add an entry into a table like so:
+-------------------------------------------------------------+
| Page | Namespace | Project | Class | Importance |
+-------------------------------------------------------------+
| First aid | 0 | Medicine | C | High |
+-------------------------------------------------------------+