Mao Fung Limthe National Institute of Human Rights Academic Fellow, shares their experiences by attending a data science meeting and workshop in September.
We started with an unexpected scene of a data science workshop on mental health – Deutsche Bank’s office in central London. This was followed by an equally surprising line from Marcos del Pozo Banos: “I don’t know how to write code in R or Python”.
This was fun and straightforward, but as I gradually discovered during the workshop – it was very realistic.
A little about myself – I’m a Fellow of the NIHR Academy of Psychiatry. I still consider myself a very Early career researcher (ECR), especially in data science. My research aims to explore the immune relationship between physical and mental health. By doing so, I turn to big data for answers and climb the steep learning curve that accompanies this quest.
I have always found programming very daunting. I attended MQ Datamind workshop to start my data science career. Although academics and the media made up most of the audience, Marcus made it accessible. Introduce programming as a basic task of communicating with a computer by numbers.
I found Marcus’ worksheet/exercise on Kaggle to be a real insight into the programmer’s mind and a useful resource to come back to in the future. The whole process was broken down into small steps, interspersed with phrases: “Now I will look for the relevant code on the Internet.” Several researchers I have spoken to since have confirmed that this is, in fact, how they are coded in R!
But of course – data science is more than just analyzing numbers and writing (or looking up the code). Even on the day of the workshop, some great examples of how NHS care and clinical research are transformed by including processes, people and expertise in existing operations (Johnny Downs and Pauline Whelan) were presented.
The next day’s conference provided more insights into data science. NIMH’s Greg Farber has asked several vital questions about the interoperability of the data we collect and efforts to harmonize it. This was also echoed by the Committee of Mental Health Funders (Wellcome Trust, MQ, NIHR), which discussed the concerted efforts of funding bodies in facilitating data interoperability and various data science initiatives such as the Wellcome Data Prize. As an ECR representative, it has been particularly fascinating to gain insight into the decision-making process of funding bodies and the (often intimidating) process for grant and fellowship applications.
Then Andrew Morris from HDR UK provided an inspiring insight into the data science infrastructure available in the UK. Referring to previous topics related to data interoperability, we have been introduced to the impact that HDR UK has had on healthcare, particularly in COVID-19, by providing huge data sets that have enabled groundbreaking research.
We were then shown how to collect data with the great examples of the MindKind (Mina Fazel) and the GLAD Study (Thalia Eley). MindKind was a great example of co-production, engaging young people in how and what data is collected and how that data is used. MindKind and GLAD were excellent examples of how to harness the reach of social media while being mindful of risks, such as potential employment aberration.
I found it fascinating to see the many applications of data science in mental health. The special highlights for me are:
- Linking Welfare and Health Records (DWP, Children in Care)
- Risk prediction models for mental health outcomes (Emmanuel Osimo, Ben Berry)
- Well-designed posters and display patterns from fellow researchers. (Special mention to Max Taquet for a very easy and accessible presentation on his findings on COVID-19 brain fog.)
Finally, I’m left with an important reminder that under the veneer of (mostly) agnostic data – it’s real life. As Ann John said while introducing John Niven, we must remember that in data science, we use public money and data entrusted to us by real people with real lives that will be affected by the results and implementation of our research. Jon gave us a remembrance, honest, and heartbreaking account of his brother’s suicide death over a decade ago, causing tears and at least one standing ovation.
All in all, it was a great workshop and conference that far exceeded my expectations, and I highly recommend it to my fellow ECRs. Personally, I have now received a large data set that I will be working on as part of my first project. I will be taking a course on Mendelian Randomisation, which will depend a lot on my rudimentary programming skills. I am exploring national and local opportunities for data that may be useful for my future projects. I hope I have something to share at the upcoming MQ DataMind event!