Data Science Day 2024

Recap

21 October 2024

On Friday, 18 October, approximately 200 participants took part in dynamic workshops and presentations during Data Science Day, exploring the latest advancements in data science at UvA.

In virtually every academic discipline, the ability to collect, analyze, and interpret data has become essential. To highlight these developments and promote data science skills, the UvA Data Science Centre, part of the UvA Library, hosts Data Science Day each year in October.

This year, the event expanded beyond a single location, featuring multiple sessions across all UvA campuses. From the latest developments in research to hands on activities and discussions with industry representatives, it was a varied and novel programme with a lot strong input from the UvA data science community and industry professionals.

AI for Science: opportunities and challenges

The day kicked off at 10:30 with presentations by Prof. dr. Paul Groth (Scientific director Data Science Centre) and Dr. Bernd Ensing (Director FNWI AI4Science Laboratory) followed by a panel with Dr. Daniela Huppenkothen and Prof. Alfons Hoekstra on the oppertunities and challenges of AI for Science.

Presentations by Prof. dr. Paul Groth (Scientific director Data Science Centre) and Dr. Bernd Ensing (Director FNWI AI4Science Laboratory). Photo's: Vincent Vink.

Panel discussion with Dr. Daniela Huppenkothen (assistant professor in Astronomy and Data Science, staff researcher at SRON Netherlands and Data Science fellow at the University of Washington) and Prof. Alfons Hoekstra (scientific director of the new FNWI Molecular and Material Design hub).

Data science in Medicine: From ideation to production

Simultaneously, between 10:30 and 12:00, at the Amsterdam UMC, speakers and attendants explored how data science drives medical innovation from early-stage ideas and research to real-world applications.

The session at the Amsterdam UMC was organised by Dr. Athanasios Angelakis, Sr. Research Data Scientist and Assistant Professor, VUmc.com - Epidemiology and Data Science (EDS) and the Amsterdam Public Health research institute (APH) of the Amsterdam UMC and the DSC. Photo's: Jordi Wallenburg.

Presentations covered topics as Large Language Models (LLMs) in healthcare, machine learning in diagnostics and prognostics, and the role of open science in research collaboration.

Key takeaways of the session were that data science in medicine is far from trivial and that teams should be built on different expertise. Another insight was that, depending on the medical problem, med-LLMs do not always outperform general LLMs. Also, steps were identified for a successful clinical AI model: 1) Collaborate with doctors, 2) Adapt based on feedback, 3) Ensure robust QA, 4) Iterate, 5) Build trust, and 6) Keep monitoring.

Navigating the GDPR: Informing data subjects and ensuring compliance in research projects

Between 11:00 and 12:00, another morning session took place at the Bushuis (University Quarter), where participants learned everything about GDPR requirements for informing individuals when their personal data is collected from sources such as social media, corpora, archives, or newspaper articles.

Dr. Kamran Adeli (photo above) introduced the privacy protection review in the Humanities RMS portal. Dario Boot, Faculty of Humanities privacy officer (photo below), explained when and how to inform people. The session was organised by the data stewards and privacy officer of the Faculty of Humanities. Photo's: Jaap Beyleveld.

The discussion covered best practices for complying with these obligations under EU privacy law and highlighted common challenges that researchers encounter. Additionally, the Humanities RMS portal was introduced as a valuable tool for assessing GDPR compliance in research projects. This platform aids researchers in protecting privacy and ensuring they meet legal requirements while conducting their work ethically and responsibly.

Data Discovery Jam: Exploring rich social science datasets

In the afternoon, from 13:00 to 16:00, a Data Discovery Jam took place at CREA (Roeterseiland campus). Participants explored the possibilities of diverse large datasets of the faculty of Social and Behavioural Sciences.

Researchers introduced rich datasets on various topics, such as gender representation in film and television (dr. Saurabh Khanna, picture above). The session was organised by the Social & Behavioural Data Science Centre (SoBe DSC). Photo's: Jaap Beyleveld.

Participants of the Data Science Jam collaborated in a hands-on data science session, analyzing and visualizing data of their choice.

Prof. dr. Agneta Fisher presented on Emotions and populist attitudes an dr. Niels Smits on predicting student success.

Generative AI in HRM: Applications, challenges, and ethical considerations

The day concluded with a session on generative AI in HRM between 15:00 and 17:30. HR leaders, labor experts, and AI scientists debated best practices for managing Generative AI-empowered workforces, while researchers from various disciplines shared insights on the functionality and ethics of Generative AI systems.