# Grocery Data Intelligence

> Grocery Data Intelligence uses Databricks Lakehouse layers to transform raw grocery data into trusted Gold metrics, then uses an AI-ready semantic layer and action flow to help business users move from insight to action faster. Built on Databricks Free Edition.

## Pages

- [Home](/): Overview of the product, hero stats, and entry into the dashboard.
- [Executive Dashboard](/dashboard): Daily KPIs — revenue at risk, waste exposure, stockout and waste leaders, store attention scores.
- [Ask Data](/ask): Genie-style natural-language questions answered from approved Databricks Gold views via SQL Warehouse — grounded in trusted semantic business definitions.
- [Action Center](/actions): Review reorder, promotion, discount, donation, and monitoring plans before they ship.
- [Architecture](/architecture): End-to-end flow from raw grocery data through the Databricks Lakehouse, semantic layer, Genie-style Ask Data, LLM/FM reasoning, Action Center, and Lakebase-ready action state.
- [Inventory Story](/story): A plain-English daily narrative of grocery inventory risk and drivers.
- [Methodology](/methodology): Risk math, approved views, and validation status.
- [About](/about): The story, medallion architecture, and how to build it yourself on Databricks Free Edition. Includes the thank-you to Databricks leaders Denny Lee and Youssef Mrini (anchor: /about#community-feedback).

## Databricks stack

### Current
- Databricks Free Edition — build and demo without enterprise setup cost.
- Lakehouse Architecture — moves grocery data from raw to clean to business-ready.
- Bronze Layer — raw stores, products, sales, inventory.
- Silver Layer — cleaned, standardized, joined for reliable analysis.
- Gold Layer — business metrics: days of supply, stockout risk, waste risk, priority, recommended action.
- Databricks SQL / SQL Warehouse — powers app, dashboard, Ask Data + AI, and approved views.
- Gold Views — surface trusted data to the app instead of raw tables.
- Unity Catalog — catalog / schema / table / view governance.
- Semantic Layer (concept) — consistent business definitions across dashboard, Ask Data, and actions.
- AI/BI Genie-style experience — plain-English questions answered from trusted data; live for some intents today, powered by a Databricks-hosted Foundation Model.

### Next steps
- UC Metric Views — native Databricks semantic layer for reusable metrics.
- Materialized / Serving Tables — faster dashboard and AI queries as data grows.
- Lakebase — store action state: reorder plans, task status, decisions, feedback history.
- Databricks Apps — host the app closer to Databricks data and security.