simplified some of the DB connections (relying on cursor fetch iterators, so we don't load all rows into memory multiple times) so they're faster
added an intro screen with summary statistics about the interview, with a warning that things might be slow for larger interviews
added an interview where you can create fake data and delete data for an existing interview.
trying to figure out how to handle finding interviews to display the data though: generating fake data only works well (and only advised) to work on a new, local server, it can cause excess CPU usage on the server
moved the excel download file generation to be a background task if there are more than 10000 rows in the current interview stats, to prevent risk of timeout.
Fixes #9, and fixes #8.