DuckDB User Group

AI-Powered Review Analysis with DuckDB and Hugging Face: Build an Automated Customer Feedback Pipeline

*** HANDS-ON WORKSHOP, BRING YOUR LAPTOP! ***

Notes:

• Space is limited, so sign up for your spot, and note it is from 12:00 to 17:00. If you're not going to make it, please unregister.
• The workshop is on the 4th floor, rooms 04.508 and 04.509.
• Bring a photo ID for check in at reception.
• Lunch and drinks are included.

Abstract:
Ratings are slipping, thousands of reviews are piling up, and nobody knows why. In this hands-on workshop, you’ll build a pipeline that finds out automatically.

Working with 2.1 million real Amazon reviews from the Hugging Face Hub, you’ll use DuckDB to pinpoint which products are declining. Then you’ll connect that SQL to an LLM with Gradio’s `gr.Workflow`, so the model reads only the complaints that matter and reports the themes behind them. Finally, you’ll deploy it as a live Hugging Face Space, backed by Storage Buckets and callable as an API.

You’ll leave with a working app and a pattern you can apply to your own customer data!

Speaker: Pratik Patel is an AI Data Engineer and Developer Advocate at Hugging Face. He has written 3 books on programming, Cloud and OSS. An all around software and hardware nerd with experience in the healthcare, telecom, financial services, and startup sectors. He's co-organizer of Enterprise AI Atlanta and the Atlanta Java User Group, and conference chairperson for Devnexus. He also writes about his knowledge (and adventures) in tech, is a frequent speaker at tech events, and master builder of nachos.