Newsletter May 2021

Deze blogpost is onderdeel van mijn nieuwsbrief-archief. Wil je de Datakami nieuwsbrief voortaan ook ontvangen? Je kan je aanmelden via deze knop:

A consolidated newsletter about me, my PhD and Datakami.
Lots of things have happened, so here’s an update for all of you, dear newsletter readers.

I moved to Nijmegen!

In March, I moved from Enschede to Nijmegen, and I’m here to stay. The greater part of 2020, the university campus in Enschede was closed as students and employees were working from home. I was staying at my partner’s apartment in Nijmegen during the pandemic, but when the situation persisted for 12 months, we decided to start looking for a new house. We found a house really fast (coincidence? luck? good karma? techno-goblins?), so now I’m living in Nijmegen again. The best thing is that it’s right next to the local bouldering gym, so I can finish my dissertation on bouldering endorphins. The remainder of my contract with the University of Twente (6 months left!) will be done fully remotely. I’m one of those weird people who absolutely THRIVES when working from home, so I’m planning to continue doing so after my PhD.

Dissertation is at 50%

I’m about halfway with writing my dissertation! My dissertation will consist of the research papers I’ve published in the past three years, so all I have to do is write the context of that research to form a coherent narrative.
I’ve also published two new papers since the last newsletter I sent you. “Fantastic strings and where to find them” is about three new datasets with video game texts that I collected. 
For “Improving Dutch sentiment analysis in Pattern”, Lorenzo Gatti and I poked around in Pattern, a toolkit for NLP and social media analysis, to see if we could improve the sentiment analysis module (and we could!).
All my work so far is open access, so if you want to read the papers for free, you can! You can find full text copies on my personal webpage under ‘publications’:

Datakami & Sarif

Working on Datakami is going slower now that I’m writing my dissertation fulltime. However, I’m still getting together with Wouter Geraedts once a month to work on our respective businesses. Wouter is a talented embedded hardware/software engineer, and my favorite local semi-conductor & soldering expert. He’s worked on a wide range of stuff, from medical devices to industrial processes. Wouter and I go way back — Wouter was my rock climbing buddy for many years, and we even wrote a peer-reviewed research paper together during our master in computer science! If you’re looking for someone with an embedded software skillset, reach out to him via

Networking in the Dutch NLP/text mining/AI space

Currently when I’m working on Datakami, the brunt of my time is spent… talking to other people! I’ve started to reach out to people on LinkedIn that work on cool projects or at interesting companies. I’m trying to get to know the industry side of text mining, AI and NLP in the Netherlands, to supplement my academic network. So far, people have been really friendly and supportive, and I’ve learned much about how NLP is used in commercial environments. If you want to talk about how NLP is used (or: how it COULD be used) in your business, let me know! It’s been so much fun to talk about this, and I’d love to meet with all of you. 

My new hobby: investing

On a more personal side, I’ve recently become really interested in investing. I’ve been reading anything related I could get my hands on, from the history of money, trade and stocks, to fundamental and technical analysis. I can recommend “The Ascent of Money”, about the history of money, and “The Devil Take the Hindmost”, which is an overview of all speculative bubbles in history (very entertaining). I also developed a Python toolkit for analysing stocks, companies and cryptocurrencies called Hetty

Cool research at the Dutch central bank

My interest in investing is trickling down to my research work. This article in Dutch newspaper Financieel Dagblad (FD) [1] described a research project by two analysts at the Dutch central bank, who tried to predict the current economic status of the Netherlands using newspaper articles. This is very close to my interests in both investing and sentiment analysis, so I scheduled a meeting with the principal investigator (i.e. the researcher in charge) of that project. It was great to talk to him and hear about how DNB is using NLP in their macroeconomic analyses. The algorithm they used is already described in this article [2], and the researchers plan to publish a full-scale peer-reviewed paper in a journal soon (hopefully this summer). I can’t wait to read it. 

[1]  (Dutch)
[2] (Dutch)

Did you like reading this update? You can subscribe to my newsletter via this link:

Geef een reactie

Het e-mailadres wordt niet gepubliceerd. Vereiste velden zijn gemarkeerd met *