Trump’s treatment in the Opinion Section of the New York Times
Performed sentiment analysis on New York Times Op-Eds published between 2017 and 2018 and the corresponding 500,000 comments, finding that comments relating to Trump are significantly more negative. More...
Wikipedia Talk pages: Toxic and obscene?
Used different machine learning classifiers (SVM, random forest, regularised regression, neural network) to differentiate between normal, toxic and obscene comments. My final model was among the best in my year and was only marginally less accurate than that of my professor. More...
​​​​​​​
PlayerUnknown’s Battlegrounds: Is cheating contagious?
Created network simulations based on PlayerUnknown’s Battleground match data. These simulations were then used to test the hypothesis that being killed by a cheater makes someone more likely to cheat, with findings rejecting the hypothesis. More...

And the Oscar goes to...?
Exploratory data analysis with actor nominations for Academy Awards with a focus on data collection (web-scraping, APIs) and visualisation, findings showing i.a. that women tend to be significantly younger than men at the peak of their career. More...
Back to Top