In the solutions I was building, I actually had to switch to the Google BigQuery ODBC driver provided by Simba. It's a little faster (albeit still not ideal). I did hit some bumps in it though. Alteryx doesn't appear to correctly translate data types in every case and I had to explicitly set a few numeric ones with CAST when pulling from BQ. It's worth noting, that when I had the data type issues, I opened a case with support and they told me that the only way it was supported was to use the BQ tool, even after I showed them it wasn't feasible.