“I’m a statistician with a background in the social sciences, but health data, and specifically dementia data, is a relatively new context for me to work in. As part of ROADMAP, I’m working with the ELSA, UK Biobank and Whitehall datasets. I’m aware that DPUK works with these cohorts and I’m always very interested in new statistical methods I could use, so I was keen to go to the Next generation seminar: Analytics to find out more.
The day was structured in two halves – in the morning were presentations. I was impressed by the speakers. The great thing was how different each presentation was. For me, Andrey’s presentation on machine-learning methods for longitudinal data was particularly relevant and Brian’s topic – mixture models – was very interesting. There were over 30 of us analysts in the audience and most were, like me, based in universities. As a group we had a good lot of questions for the presenters – it was great to have that opportunity to probe the speakers and the use of the methods they introduced.
In the afternoon was the interactive group work session. Groups were made up of epidemiologists, biologists, psychologists and data analysts mainly. In these mixed groups we chose a research question and considered how we would deal with it. I’m not normally a fan of group work but I will admit that it was valuable to see how the people in my group approached the problem in different ways – sometimes very different ways – due to their different backgrounds.
For me the best thing about the day was the opportunity to learn about the new statistical approaches that are out there and how we can apply them in cross-cohort analyses. I’ve learnt about new methods which I’ll be able to draw on in future studies. That’s a great outcome for me and means I’ll definitely be on the look-out for other events like this.”