Notes on the RSS One-Day Conference, by Rosemary Tate and Skye Marshall
Bayes Centre, Edinburgh | 22nd October 2026
Aims
With the advent of AI, data quality has never been more relevant to organisations all over the world. While most readily admit they have a serious problem, beyond short-term “data cleaning,” they simply do not know what to do about it!
This provides an enormous opportunity for practicing and research statisticians! After all, many of our approaches and tools can help find and eliminate sources of error, evaluate whether a collection of data is fit-for-use in addressing the problem at hand, and quantify results.
The aim of this (very timely) conferences was to help form a community of practitioners and researchers with an interest in seizing this opportunity!
Talks
The format was a little different from most events run by the RSS. Along with our three keynote speakers, we invited members to submit abstracts for talks. We received 16 submissions in total, and out of these, 12 were selected for presentation, either as a 25-minute talk or a 10-minute presentation as part of a panel.
The resulting talks proved to be both complementary and varied. The first session was chaired by Ruth King, director of the Bayes Centre, who kept everyone strictly to time. With so many questions from the audience, she had to limit the Q&A and encourage attendees to continue the lively discussions during the breaks.
Details of the programme and talks can be found here.
The afternoon featured two panel sessions moderated by Tess Vernon, Chair of the local Edinburgh RSS section. Both sessions generated excellent discussions and prompted a wide range of questions for the panelists.
We finished the day with a group discussion
“All together: Building a community”. Notes from the discussion are shown below.
Community Discussion
Aim: How to build a community and to change the world in relation to data quality.
General strategic direction suggestions
- Spreading the joy and fun of getting quality data.
- Making data quality a fashionable term/buzzword so that organisations include it in their annual strategic direction documents. This would make them more likely to respond to suggestions such as implementing data quality within their journal guidelines for example.
- Could include a slogan like ‘Quality Data Saves Lives’
- Make data quality an antidote to the increasing feeling of having no clarity or control over information (e.g. in the age of AI and fake news). Allowing people to find trust by providing a framework for truth seeking.
- Letting other people do truth for us - now let's do it ourselves.
- Get in touch with the media.
Suggestions on the project-level
- Curricula & Examples: Add data quality to school and higher education curricula and provide good examples.
- Analysis Plans: Put how data quality was checked into Statistical Analysis Plans.
- Target Audience: Target scientists rather than statisticians/programmers by posting in Significance magazine.
- Website: Build a website.
- Book - produce a multi-author book
- Case Study: NHS Palantir as a case study.
- RSS Support: Get support from the RSS. Contact the RSS ambassadors.
Suggestions on pulling together a group to do the work
- Make a LinkedIn group.
- Connect everyone who was at the event on LinkedIn and make sure we stay on each other’s algorithms for the next few months by:
- Posting about the event (maybe a photo) and tagging everyone who was there.
- If everyone then interacts with that post (likes and comments and shares) then we are more likely to be ‘organically’ connected.
The key message was to make data quality fun and cool.
Slides from the event can be found here.