Data scientist with a versatile background and a strong interest in social and behavioural sciences, finance and artificial intelligence. Experience with project management, data dashboards, machine learning and academic writing. Currently in the last year of my MSc. Following courses in mathematical statistics, multivariate statistics, data mining and statistical learning.
Part of academic team investigating the use of Cortical’s ‘Semantic Folding’ in predicting stock and commodity price volatility using close to one billion news articles going back to 2016. Responsibilities include:
- Building and maintaining infrastructure using Python, PostGreSQL, Docker and Google Cloud Services to process historical and current news articles using Cortical’s semantic folding algorithm.
- Aid in collecting and preprocessing training data to create custom models using Cortical’s semantic folding algorithm.
- Design and implement downstream tests to assess the usefulness of custom models, using e.g. monte carlo simulations, clustering methods and convolutional/recurrent neural networks.
- Provide support to other researchers in terms of statistical expertise and by creating custom docker applications that simplify the access to the custom models.
Investigated the application of Cortical.io’s core technologies to finance & investing. Focused on:
- Gaining an in-depth understanding of Cortical.io’s core technology and its application to large, textual datasets.
- Develop methods to process, store and query news articles using Cortical.io’s semantic folding technology
- Use these methods to develop a prototype for visualizing and downloading historical and live data representing the nature and quantity of news relevant to user-selected stocks, commodities and portfolios.
- Serving as liaision between the research team led by Prof. David Stolin and Cortical.
The Center for Innovation facilitates the development of Massive Open Online Courses (MOOCs) at Leiden University. I was responsible for retrieving, storing and analysing MOOC data with the aim of improving courses and providing feedback to content creators & academics. Co-published several academic articles with researchers and published open-source tools to process and analyze MOOC data.
The HCSS helps governments, non-governmental organisations and the private sector to anticipate the challenges of the future with practical policy solutions and advice. While at the HCSS, I worked on several projects in areas of Big Data, security studies, and international development. Further set up an event-monitoring server that scraped and processed roughly 20.000-50.000 news articles each day, and created a package in R to analyze assertive behaviour among ‘great powers’ such as Russia, the United States and China.