Showcase

Predicting Covid-19 outbreaks using multi-layer centrality measures

Preprint

View on arXiv | View on Github

In this study, we leverage the framework of multi-layer networks to model and predict the spread of COVID-19 in the Dutch population. We compare different centrality measures in their predictive ability of individual infections. The centrality measures are employed as predictors in proportional-hazards and XGBoost models, and we find that Eigenvector centrality can account for substantial variation in infection risks and timing. The study originated from my master thesis at the Departmet of Methodology and Statistics at Utrecht University.

Master Thesis

View on Github

Mapping Missions: New Data for the Study of African History


This project originated from my bachelor thesis in Political Science at University of Bremen. After digitizing a new map on the locations of Christian mission stations in Africa, we investigated how this map differs from the commonly used sources in historical social research. You can find a synthesis of the findings in this Twitter thread.

CorrelAid Workshop: Tidy Network Analysis in R and Python

View on Github


As a data science volunteer for the NGO CorrelAid , I gave a workshop on network analysis in R (and Python) to my peers. I focused on packages and workflows from the tidyverse, and also implemented the workshop presentation itself using rmarkdown.

Co-authorship network of the SOCIUM research center

Data collection

View on Github

During my Bachelor’s at the University of Bremen, I assembled a dataset of co-authorship relations within the local social science institute SOCIUM.

Working paper

View on Github


As part of a summer school course at Radboud University Nijmegen, I returned to this dataset for the implementation of a temporal exponential random graph model (TERGM). I try to explain the temporal emergence of the co-authorhsip network over 10 years, using author attributes and internal network measures.

TidyTuesday


I have participated in several TidyTuesday events to experiment with visualization in ggplot2. View here, here, and here .