A 1.5-hour hands-on workshop where you build an end-to-end AI analytics solution entirely within Snowflake — no external APIs, no additional tools, no data leaving your account.

What You Will Build

What	How
Synthetic lottery data	10,000 players, 24 months of draws, 30 charity partners
AI-enriched player intelligence	Sentiment analysis, segmentation, insights extraction, and personalized messages using Cortex AI
Interactive Streamlit dashboard	Branded with Postcode Loterij colors — player KPIs, charts, live winner draft with animation, AI chatbot, and AI player scoring
Semantic View & Cortex Agent	A business-friendly data model that powers a natural language AI agent

Architecture

                       SNOWFLAKE ACCOUNT
    ┌─────────────────────────────────────────────────────┐
    │                                                     │
    │  ┌─────────┐   Cortex AI    ┌──────────────┐       │
    │  │ RAW     │ ────────────>  │ ANALYTICS    │       │
    │  │ Schema  │  Sentiment     │ Schema       │       │
    │  │         │  Classify      │              │       │
    │  │ PLAYERS │  Extract       │ PLAYER_      │       │
    │  │ DRAWS   │                │ INTELLIGENCE │       │
    │  │ TICKETS │                │              │       │
    │  │ CHARITIES│               │ SEGMENT_     │       │
    │  │ DONATIONS│               │ SUMMARY      │       │
    │  └─────────┘                │ CHARITY_     │       │
    │                             │ IMPACT       │       │
    │                             │ DRAW_RESULTS │       │
    │                             └──────┬───────┘       │
    │                          ┌─────────▼──────────┐    │
    │                          │ SEMANTIC VIEW       │    │
    │                          │ Business data model │    │
    │                          └─────────┬──────────┘    │
    │                    ┌───────────────┼────────┐      │
    │            ┌───────▼────────┐  ┌───▼──────┐ │      │
    │            │ STREAMLIT APP  │  │ CORTEX   │ │      │
    │            │ Dashboard      │  │ AGENT    │ │      │
    │            │ Winner Draft   │  │ NL → SQL │ │      │
    │            │ AI Scoring     │  └──────────┘ │      │
    │            │ AI Assistant   │               │      │
    │            └────────────────┘               │      │
    │                                             │      │
    └─────────────────────────────────────────────────────┘

Snowflake Features Covered

Feature	What it does	Module
GENERATOR + UNIFORM + CHR	Generate realistic synthetic data entirely in SQL	1
CORTEX.SENTIMENT	Score player feedback from -1 (negative) to +1 (positive)	2
CORTEX.AI_CLASSIFY	Categorize players into segments (High-Value Loyal, At-Risk, etc.)	2
CORTEX.AI_EXTRACT	Pull structured fields (interests, complaints, preferences) from free-text feedback	2
CORTEX.SUMMARIZE	Generate concise summaries of player feedback	2
CORTEX.COMPLETE	Generate personalized retention messages and real-time AI player scoring	2, 3
Streamlit in Snowflake	Build interactive dashboards without leaving Snowflake	3
Semantic View	Define business meaning of tables (dimensions, facts, metrics) for AI	5
Cortex Agent	AI orchestrator that converts natural language to SQL via your Semantic View	5
Container Runtime	Run Streamlit on a compute pool with REST API access (SPCS)	5

Prerequisites

• A Snowflake account with ACCOUNTADMIN access
• A web browser (everything runs inside Snowflake — no local installs needed)

Lab Agenda

Module	Topic	Duration
0	Environment Setup	5 min
1	Foundation: Create Database & Synthetic Data	15 min
2	AI Enrichment with Cortex AI Functions	25 min
3	Build Streamlit Dashboard	20 min
4	Explore & Interact with the App	10 min
5	Semantic View & Cortex Agent	20 min
Bonus	Cortex Code Challenge	15 min

Getting Started

1. Open the Lab Guide — this is the step-by-step walkthrough for the entire workshop
2. Start at Module 0 to set up your environment
3. Follow each module in order — every step explains what you're doing and why

Repository Contents

postcode_loterij_lab/
├── README.md                 ← You are here
├── lab_guide.md              ← Full step-by-step lab guide (start here)
├── assets/
│   ├── banner.svg            ← GitHub README banner
│   ├── border.svg            ← Red-to-blue gradient bar
│   └── divider.svg           ← Diamond section divider
├── scripts/
│   ├── 00_complete_worksheet.sql ← All SQL in one importable file (facilitator backup)
│   ├── 01_setup.sql          ← Module 1: Database, tables, synthetic data
│   └── 02_ai_enrichment.sql  ← Module 2: Cortex AI enrichment pipeline
├── streamlit_app.py          ← Module 5 final version (includes Agent chatbot + AI Scoring — do NOT use for Module 3)
├── requirements.txt          ← Module 5: Container runtime dependencies (PyPI)
└── environment.yml           ← Module 3: Warehouse runtime dependencies (Conda)

Warning: The streamlit_app.py file in this repo is the Module 5 final version with the Cortex Agent chatbot and AI Player Scoring tab. It requires container runtime (compute pool) and will not work in Module 3's warehouse runtime. When following the lab guide, use the code provided in Module 3, Step 3.3 — do not copy from streamlit_app.py until you reach Module 5.

Note on environment.yml vs requirements.txt: The lab uses two different dependency files at different stages. In Module 3, the Streamlit app runs on warehouse runtime (Conda-based), which uses environment.yml. In Module 5, the app is upgraded to container runtime (PyPI-based), which uses requirements.txt. Both files are included in this repo for reference. When following the lab guide, you'll switch from one to the other at the appropriate step.