I decided to create a fun Formula 1 demo dataset for showcasing the integration between Fabric Data Agents & Microsoft Copilot and/or AI Foundry. There are a lot of datasets out there and this one is inspired by the initial work done by Shubham Yadav a few years ago. Unfortunately the data source he references is no longer available. Thus, I did some searching and found the kaggle - Formula 1 Race Data. This dataset is updated after each F1 race and using the solution I created in this repo, you can easily update the medallion architecture on a weekly basis during race season.
To get started, please perform the following:
- Create a brand new Fabric Workspace. I named mine F1.
- Clone this repo to your GitHub environment.
- Add your newly cloned GitHub repo to your Fabric Workspace environment via the Git integration under Workspace settings. For more on this topic, check out Connect a workspace to a Git repo - GitHub Connect.
- Within your Fabric Workspace, perform an Update under Source Control to pull the Lakehouse, Notebooks and Data Agent into your newly created workspace. For more on this topic, check out Basic concepts in Git integration - Commits and updates.
This repository contains the following items that will be deposited into your workspace once you perform the Update via Source Control referenced above.
- F1 Lakehouse with schema enabled
- Silver - All Tables: This notebook downloads the .csv files to the bronze file path from Kaggle and transforms them into delta tables in the Silver catalog. There are two variables to update in the notebook to point to the information for your environment
- bronze_file_path - Set this to match your bronze file path for you will deposit the Kaggle CSV files.
- silver_schema - Set this to match your LAKEHOUSE.SCHEMA for your silver delta tables.
- Gold - All Tables: This notebook transforms all the tables into the Silver schema to delta tables that act as materialized views in the Gold schema. There are two variables to update in the notebook to point to the information for your environment.
- silver_schema - Set this to match your LAKEHOUSE.SCHEMA for your silver delta tables.
- gold_schema - Set this to match your LAKEHOUSE.SCHEMA for your gold delta tables.
- f1_agent: This is a Fabric Data Agent that utilizes the tables in the f1.gold schema as it's data source. It also includes comprehensive Data Source Description, Data Source Instructions, and Agent Instructions.
The beginning cells of the Silver - All Tables notebook download the following .csv files to your bronze file path from the kaggle - Formula 1 Race Data web site.
- circuits.csv
- constructor_results.csv
- constructor_standings.csv
- constructors.csv
- driver_standings.csv
- drivers.csv
- lap_times.csv
- pit_stops.csv
- qualifying.csv
- races.csv
- results.csv
- seasons.csv
- sprint_results.csv
- status.csv
For this repo, I created the following items in my F1 lakehouse. The notebooks and the f1_agent data agent references are based on these items.
- I created a bronze folder in my F1 Lakehouse Files and uploaded the csv files to this location.
- I created 2 schemas under the lakehouse:
a. silver
b. gold
Below are some sample suggested prompts/sample questions we used to test the Fabric Data Agent and to demonstrate the integration within Copilot Studio and AI Foundry.
- What are the constructor standings for 2025?
- Which drivers finished on the podium in the last race?
- Who won the United States Grand Prix in 2025?
- Which drivers have the most wins this year?
- Which drivers have the most wins in the last 10 years?
- What are the constructor champions for the past 20 years?
You can also integrate this Fabric Data Agent with both AI Foundry and Copilot Studio. For more on this topic check out the following articles:

