22-25 April 2026

Sampling Bias: When Data Collection Shapes the Story

Proposed session for SQLBits 2026

TL; DR

This session explores how sampling bias subtly shapes data narratives. Through real-world examples, it highlights how design choices can obscure or reveal gaps, urging ethical, transparent communication in data storytelling.

Session Details

Sampling bias is rarely obvious, and that’s exactly why it matters.

In this lightening talk , we’ll explore how limitations in data collection and sampling choices can shape the story an audience takes away, even when the analysis and visualisation are technically correct.

Using real-world examples, I’ll show how certain populations can be unintentionally over-represented, under-represented, or excluded entirely, and how visual design decisions can either mask these gaps or make them transparent.

This session focuses on awareness rather than accusation, understanding the limits of what a sample can tell us, and making those limits clearer through thoughtful data communication and visual design.

3 things you'll get out of this session

- Understand what is sampling bias and why it matters even when we don’t collect the data ourselves - Recognise how sampling bias influences the conclusions audiences draw from data - Apply techniques to communicate data limitations clearly, ethically, and without assigning blame