Projects from my master's and my own independent work, grouped by where they stand. The code for most of them is public on GitHub.
Supervised research internship at the Institute for Analytical Sociology. Social scientists often draw distances between respondents, parties or countries as flat two-dimensional maps, and those maps distort. I build distance matrices from survey data, place them on the surface of a sphere and compare the fit with standard PCA and MDS using Kruskal stress.
An independent secondary analysis of which socioeconomic and experiential factors go together with positive views of AI at work. The data are cross-sectional, so the results describe associations, not causes.
How do Nordic universities frame generative AI in teaching? The pipeline collects official guidance, processes it and codes it on two independent indices: a Restriction and Enforcement Index and a Pedagogical Integration and Support Index. A policy can score high on both. The pilot covers one university in each Nordic country and is built to scale to 30 or more European universities.
A study design on how socioeconomic and demographic background predicts AI use, trust and learning behaviour among university students in Sweden, framed as a question of digital inequality. The proposal and a 38-item survey instrument are written.
How do best-friend networks change over a school year? A longitudinal analysis of one secondary-school class across three waves, testing reciprocity, same-sex preference and triadic closure.
Do personal and economic freedom form one shared structure across countries, or distinct country profiles? Eleven indicators for 161 countries, reduced with PCA and grouped with hierarchical clustering.
Three models of how large patterns emerge from local interactions: inequality under random exchange versus redistribution, diffusion with social blacklisting across network types, and a tipping point in a forest fire model.
A scraper that turns the public Visit Stockholm events pages into a structured dataset of events by category, date and location.
A tutoring web app that answers in the language of the learner's device (Portuguese, English, Spanish or French) and keeps the conversation on educational topics.