22-25 April 2026 
ChatGPT for SQL People: An Introduction to LLM-Powered Data Analysis
Proposed session for SQLBits 2026TL; DR
A pragmatic 20-minute intro to using ChatGPT and the OpenAI Python library for data analysis: prompt patterns for text, simple natural-language-to-SQL, and a small Python workflow you can take home and adapt.
Session Details
Large language models are rapidly becoming part of the everyday toolbox for data professionals – not as a replacement for SQL, but as a way to analyse data that doesn’t fit neatly into tables.
In this 20-minute session, we’ll look at how you can use ChatGPT and the OpenAI Python library to solve real analysis tasks that most SQL and BI people recognise: classifying and extracting information from text, and turning natural-language questions into SQL that can be reviewed and executed safely.
We start in the ChatGPT UI with a few simple prompt patterns for text classification, extraction and summarisation. Then we turn the same ideas into a small, repeatable Python workflow that calls the OpenAI API – something you can take home and adapt to your own environment.
Along the way we’ll discuss where LLMs shine, where they fail, and which basic guardrails you should add before letting a model anywhere near your database.
The goal is a pragmatic, realistic introduction to LLM-powered data analysis for SQL professionals – grounded in real-world scenarios rather than hype, and focused on what you can do today without rebuilding your entire data platform.
In this 20-minute session, we’ll look at how you can use ChatGPT and the OpenAI Python library to solve real analysis tasks that most SQL and BI people recognise: classifying and extracting information from text, and turning natural-language questions into SQL that can be reviewed and executed safely.
We start in the ChatGPT UI with a few simple prompt patterns for text classification, extraction and summarisation. Then we turn the same ideas into a small, repeatable Python workflow that calls the OpenAI API – something you can take home and adapt to your own environment.
Along the way we’ll discuss where LLMs shine, where they fail, and which basic guardrails you should add before letting a model anywhere near your database.
The goal is a pragmatic, realistic introduction to LLM-powered data analysis for SQL professionals – grounded in real-world scenarios rather than hype, and focused on what you can do today without rebuilding your entire data platform.
3 things you'll get out of this session
Understand when LLMs are a good fit for data analysis tasks alongside SQL
Apply three concrete prompt patterns in ChatGPT for text classification, extraction and summarisation
Implement a small Python workflow using the OpenAI library to automate the same analyses
Speakers
Stephan Stoltze's previous sessions
Master Data Services 2016
In this session you will get a brief introduction to the basic principles of Master Data Services, together with an overview of all the new features brought to you with Master Data Services 2016.
Real Life Master Data Management with SQL Server and MDS
In this session you will learn about common MDM scenarios, and get Real Life examples on how to handle these Scenarios by building a MDM Architecture using SQL Server and Master Data Services.
Excel 2013 - Whats new beside PowerPivot and Power View?
You probably already heard a lot of the wonders of PowerPivot and Power View, but will you have any other cool new BI features with Excel 2013. The answer is a big YES! This session will demo a lot of those.
Business Intelligence with Excel 2010
Most of us use Excel in our Daily work, but are you familiar with the Business Intelligence features in Excel? In this session we'll look at some of the most common and useful BI features in Excel 2010 and discuss limitations and best practices.
Writeback-Here Comes the Sun
With Excel 2010 it is finally possible to modify data in a SSAS cube, without using any macros or third party add-ins. This opens for a lot of common business scenarios like budgeting and planning directly from Excel 2010.