My product (https://gitsense.com) moves most of Git's history into a relational database and then I do indexing on top of that to produce things like code churn (lines added, changed and deleted), which is also stored in the database. So as you can probably imagine, generating the table is just a bunch of SQL statements.