My workflow usually reads something like this: 1. Import data using hand-written SQL from SQL server (read_sql) 2. Perform various filtering, aggregation, mat

My workflow usually reads something like this: 1. Import data using hand-written SQL from SQL server (read_sql) 2. Perform various filtering, aggregation, math (loc, iterrows/itertuples, groupby, agg) 3. Push results either out to delivery file or back to source server (to_csv, to_excel, to_sql). Really bread and butter stuff, but the relative ease and stability of using pandas to do it is the attraction.