Creating Semantic Models for Snowflake Cortex Analyst

Learn how to generate your own semantic model from your Snowflake data and leverage existing partner semantic models for use with Cortex Analyst.

Link to demo video Link to Quickstart
Solution Innovation

overview

Snowflake has introduced Cortex Analyst, a cutting-edge solution designed to revolutionize how business users interact with their data. Currently in Public Preview, Cortex Analyst is exposed as a REST API and offers a conversational interface that allows users to explore structured data with ease. Unlike traditional BI tools that often require complex queries or reliance on data teams, Cortex Analyst empowers users to ask questions in natural language and receive accurate, real-time insights.

To get started with Cortex Analyst, one only needs a semantic data model in the form of a YAML. Developers at Snowflake, in conjunction with the Snowflake Solution Innovation Team, have developed an open-source Semantic Model Generator using Streamlit to assist you in taking your semantic model from zero to one.

Snowflake Features Used
Quickstart
Build this solution now Creating Semantic Models for Snowflake’s Cortex Analyst

Solution Architecture: Creating Semantic Models for Snowflake’s Cortex Analyst

  • The Semantic Model Generator deploys a local Streamlit front end to guide users through automated generation of a Semantic Model yaml file for Cortex Analyst.
  • The Generator analyzes selected tables for columns, data types, sample data and employs Cortex Search for table and column descriptions.
  • The Partner Translator analyzes partner semantic models and merges relevant partner metadata with metadata from the Generator into the Cortex Analyst Semantic File format.

HANDS-ON TECHNICAL RESOURCES