Sync user IDs across all data sources - regenerated test_candidates_demo.csv with users [3,12,15,20,25] that exist in both orders and candidates

Initial deployment