Using Postgres to locate the best coffee near you!
summary
A practical walkthrough that turns a Zomato dump of 8,000 businesses and 50,000 reviews into a working search: JSON ingestion into Postgres, TSVector full‑text, embeddings via Azure AI into a Postgres vector column, DISCAN vector index and PostGIS geometry, then a hybrid query (cosine nearest‑neighbor + spatial filter) for 5 km radius results. Demo shows ~20–30 candidate cafes fed into a Python RAG application that narrows choices by diet, walk time and takeaway.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker intro
- 3:14Dora’s map for coffee search
- 5:00Why Postgres for coffee discovery
- 8:52Pattern matching search limits
- 10:40Full text search in Postgres
- 14:31Understanding vector embeddings
- 20:54Semantic search with vectors
- 26:58Adding geospatial data
- 29:24Hybrid search combining vectors and location
- 31:56Retrieval augmented generation demo
- 37:45Live RAG application walkthrough
- 40:11Building AI apps with Postgres
- 41:36Q&A and closing
related talks
48mUsing Postgres to locate the best coffee near you!
42mPostgreSQL Tips & Tricks For App Devs
44mPostgres and OpenAI - a match made in Azure
28mAdvancing Drug Search with PostgreSQL and Azure AI
05mPostGIS & the Hard Rock Café collection
43mFull-Text-Search Explained from A to Z with French Food
33mPostgres, the Best Tool You're Already Using
29mA Pythonic semantic search