‘German? I’m European!’

Even before the current movement, I’ve often asked myself what my country of origin, Germany, has learned from the Third Reich. Has the past given us some level of immunity for the future? I’ve got my doubts, but also think there’s hope.


Do-Not-Track is dead. Long live Do-Not-Track!

Tracking, the collection of data about user behaviour, is widespread on the web. For this reason, the idea of a “Do-Not-Track” (DNT) setting emerged a little more than a decade ago, in 2009.

This system gives users a simple choice to reject and accept online tracking. Unfortunately, DNT failed, and tracking continues.

I explore the reasons for this failure, but also how DNT has managed to change user privacy for the better, on the blog of the HCC group at the University of Oxford.


Perceptions of YouTube’s political influence

YouTube plays an increasingly important role as a political medium. Yet, the implications are to-date not well understood and difficult to analyse, since access to YouTube’s statistics is limited.

To address this gap, my friend Yury Kolotaev and I surveyed 124 people about their views and experiences around YouTube’s political influence.

Our results reveal diverse, sometimes conflicting views on YouTube’s growing political role, and highlight the need for more research, discussion and possibly regulation.

Read our full account on Medium.


Birthdays in quarantine

Today, it’s my birthday and I’m spending it by myself, self-isolating. And that’s alright. In fact, I love it.

The whole world out there seems to be going crazy these days. But, time will come. Things will work out again.

It’s interesting to take the crisis as an opportunity. It shows us how fragile and endless life is. What does really matter?

Living abroad, I don’t usually spent much time with my friends and friend from home. So, I wanted to make my birthday a special occasion and have a proper get together for my birthday, my first birthday party in 4 years.

I didn’t have a party, but it’s been special notwithstanding.

My family reached out to me on FaceTime, and so did many of my friends.

Whilst spatially apart, I’ve never felt this emotionally to my loved ones.

We’re finally taking time for another again, rather than always chasing something, prestige, success, whatever..

But, in times of crisis, this all doesn’t seem to matter anymore. What matters is the foundation of our existence.

I am full of hope that the crisis will end up changing society for the better.

This hope is likewise fear. Once the wind has settled, many things will quickly return to normality.

Let’s stay foolish.


LDA, Probabilistic Topic Modelling and RMarkdown

This week, I started off learning about RMarkdown. RMarkdown allows you to combine code and text. Pretty cool! You can analyse data in R or Python, and then use the results as variables in your text. This text can easily be compiled to various formats, including pdf, docx, and html. Hence, this article, too, was written in RMarkdown. I could upload it with ease. Find the pdf version here:

Probabilistic topic modelling is a method to categorise a given set of documents, that is a collection of texts. Such a set of documents could be a collection of news articles. Every news article would then be assigned a topic. A topic is a group of documents that use similar words. An example of a topic could be “parliament”, “government”, “spending”, so a politics topic.