• In general, you're going to have a hard time quantifying this precisely. The issue is that you're relying on one of two pieces of information, the login and the application in the connection string. The problem with the login is that so many people and applications share a single login, sometimes even the 'sa' login. That makes it difficult to differentiate queries from say, the DBA and a given application if both are using the 'sa' login. The application value in the connection string is completely optional. Some will fill it in. Some won't.

    But, what you can do to try for this is capture the query metrics (I'd probably use Extended Events, but you can use Trace) and then generate reports based on login and application, but with the understanding that these are likely to be less than completely accurate.

    You can also toss in filters by database which is likely to differentiate by application too.

    "The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
    - Theodore Roosevelt

    Author of:
    SQL Server Execution Plans
    SQL Server Query Performance Tuning