Till startsida
To content Read more about how we use cookies on gu.se

From Psychology to Applied Data Science

The master's programme in Applied Data Science started in 2017 and Fionn Delahunty from Ireland is in the first batch of students in the programme. The students are from all over the world and several of them have like Fionn chosen to combine a completely different subject with data science.

How would you describe your programme?

– The applied data science programme is quite well suited to those from non-traditional backgrounds, personally I have a bachelor’s in psychology from Ireland. I think the course is perfectly suitable to upskill on my existing knowledge and give me the tools to tackle bigger and more interesting problems. Some people have asked why I changed from psychology to data science, personally however I don’t see it as a change. Data science is part of everything today and I still do research in psychology, but I can now do some better and more interesting research, because of the skills I have obtained from this field.

Why did you choose this specific master’s education?

– Data science is a hot topic right now, almost every university seems to be setting up master programs related to it. Most of those programmes seem to be aimed for computer scientists. Coming from psychology there where limited options for me. Only one or two other courses in Europe. It’s seems a bit bizarre, given that much of the advancement in data science and machine learning is in applied and novel areas.

– This course was exactly what I was looking for, 60 percent of the time spent learning about core data science and machine learning topics, and the remaining time I could choose specific courses that I felt would give me skills for my future careers.

– I think having a background in human behaviour also gives me an interesting view on some aspects of data as well. When you’re dealing with behavioural big data, there’s a temptation to forget that you’re actually looking into people's lives, and your decisions may end up affecting people's lives.

How did you realise you wanted to combine psychology and data science?

– Toward the end of my bachelor’s degree I started to see research coming from Facebook. The number of participants was just under a million, making it one of the largest scientific studies in history. Facebook divided the users into three groups and changed what the users saw on their profile. In one of the groups the users primarily saw sad posts from their friends, one group saw primarily happy posts and one group was a control. The people who got sad posts started to post sad things themselves, and vice versa for the group that got happy posts. The experiment successfully tested the social contagion theory. That a person behaviour is influenced by their social group.

– Although it was a simple experiment, it was basically the largest ever behavioural randomised control experiment performed in history. I saw that this was the future, the days of bringing small numbers of people in to labs to participate in experiments where gone.

– There were of course major ethical concerns with this study, although people gave consent in the terms and conditions when they joined Facebook, they didn’t really give informed consent to be part of the study. Legally yes, but not in the way we’d view it in psychology or medicine. The bigger issue really is what Facebook is doing and not telling us, they’re not a public company and they don’t answer to any ethic committee. Europe is trying to put laws in place to protect people, but fundamentally people still don’t accept that companies could be changing their offline behaviour without their knowledge. Until people realise that, like when the population realised smoking was harmful for your health, things won’t really change.

Fionn recommends an article in Technology Review – "What Facebook knows"

So you think there is an increasing need for ethical regulations?

– Yes. If you want to design a clinical drug for example, you have very strict rules on clinical trials. We often hear about companies making mistakes and people can end up in jail. There are no real laws or knowledge about tech companies doing the same. If they do an experiment and something goes wrong, can anyone be held accountable?

– My research at the moment is about detecting depression in social media using machine learning. The project is interesting, and my research group wants to help people, but we also think our research might be used by other people with other intentions. However, we feel that public peer review process in Academia is important. It’s ensures a certain quality of work, better than might be seen inside a private company.

– I suspect insurance companies and HR companies are already using prediction systems like this. The difference in academia is that every decision I make in my research must be backed up by science. My work is reviewed by people in AI and medicine who ensure it’s the best it can be. A data science team in an insurance company doesn't have that oversight. They can do pretty much what they want and build models that might incorrectly predict a certain disorder. When you’re working with people lives, I think you need the highest level of scrutiny.

- I’m entering this field right at a serious time of change, laws like GDPR are coming into effect as we speak. In some cases, we are the first people to try and test these laws, and that can be interesting. It really does feel like you’re on the edge of something.

Has the master's programme met your expectations so far?

– Data science is an interesting area since there are so many fields it can apply to. The concept of what data science is hasn’t even been fully defined yet. I didn’t have many expectations about the programme. In my summer internship after the first year, I looked at my colleagues who graduated from other universities and I feel confident that my skills are on par, or higher in some areas than theirs.

– It was strange to change from a bachelor’s in psychology to a master’s in data science and it has been a tough year. Some of my favourite courses are Applied Machine Learning and Techniques for Large Data Sets. I found out when I started working that these courses are especially useful, more so than I thought when I was studying them. I also found some courses in the maths department regarding more traditional statistics to be quite useful.

– About one quarter of the students in the programme are from Sweden originally, so you get to know and work with people from all over the world which is a major advantage I didn’t expected. As a result, you develop quite a lot of important soft skills.

What is it like to be a student in Gothenburg?

Naturally there are a couple of issues, housing being the major concern. Despite that, it’s a great city to live in. Not too big, not too small, friendly people and lots of nature throughout the city.

What about the future?

– I plan on continuing my research in depression detection for my thesis. Afterwards I hope to move into a career, and hopefully return to academia to complete a PhD in the future.


Text: Catharina Jerkbrant
Photo: Simon Ungman Hain

Fionn Delahunty

Fionn DelahuntyFionn Delahunty from Ireland found a master's programme in data science that suited him as a student with a BSc degree in psychology.

Page Manager: Simon Ungman Hain|Last update: 3/7/2019

The University of Gothenburg uses cookies to provide you with the best possible user experience. By continuing on this website, you approve of our use of cookies.  What are cookies?

Denna text är utskriven från följande webbsida:
Utskriftsdatum: 2020-08-08