Who are the men and women behind the data science assignments at Ordina? Drawn from a wide range of backgrounds, they have one thing in common: they are all bright minds. Take David De Wachter. At the end of 2017, David swapped his academic career for a position as a Data Scientist in the VisionWorks team. ‘It was time to actually put theory into put practice,’ he explains.
Growing up, David’s dream was to become a policeman. But at the age of 18, he decided to go to VUB, Vrije Universiteit Brussel, where he studied a degree course somewhat related to the job of policeman: sociology. ‘The perfect choice of degree course when you are not quite sure yet which way you want to go,’ David opens. ‘A year into the course, it was clear to me that I was mostly interested in practical subjects like research methods, applied statistics and demography. I built on this.’ David was that passionate about quantitative sociology that he decided to take a PhD, which resulted in a double PhD at VUB and the University of Antwerp. Even after this achievement, he continued his university studies, this time as a PhD researcher at the University of Antwerp.
Big data before the term was coined
Type his name into Google and you will find a whole bunch of research projects David was involved in. ‘Using millions of records from Belgian censuses, the National Register etc, we solved demographic problems. That was actually big data before the term was coined,’ he says. ‘And extremely interesting. But after nine years of research, I was ready for a change. Research work can sometimes be a theoretical and solitary business. I wanted more interaction: to work in a team, developing solutions for c0ncrete problems. What’s more, I wanted to broaden my horizons: as long as you are working in one particular field – in my case, demography– you are mostly using a series of specific techniques. I was keen to explore new things.’
The millions of records from Belgian censuses, the National Register etc we used to solve demographic problems, that was actually big data before the term was coined.
Research and people
In 2016, David left university to work at IFIC, the Instituut voor Functieclassificatie – a non-profit organisation engaged in occupational classification and related activities – where he calculated the cost of new pay scales. ‘In addition to being a data analyst, I also provided technical support at negotiations between employer and employee organisations. Quite intense, but also great fun,’ David recalls. ‘It reinforced a feeling I had at university, that I love research and technical things, but that I also really love working with people. When the project was completed, there was no doubt in my mind: I was going to be a consultant.’
I love research and technical things, but I also really love working with people. A job as a consultant allows me to do both.
Programming, analysing and visualising
‘In October 2017, David joined the VisionWorks team at Ordina as a Data Science consultant. ‘Initially, I was unsure whether I had the right skills. I had a background in research and applied statistics, but Data Science also involves programming and IT. I didn’t feel quite skilled enough yet in those areas. Ordina won me over.’
During his first months at Ordina, David worked on an internal tool to automatically assign vacancies to the right business unit, based on Natural Language Processing (NLP). Soon after that, he was assigned to FSMA, the regulatory body of the Belgian financial market. FSMA had created a number of theoretical risk models for financial products and called in Ordina to help programme – and consequently, automate – these models and to optimise them where possible. This project also involved quite a lot of data visualisation: David built a dashboard for the experts, which they could also run simulations in.
Predicting time series at UCB Pharma
A number of VisionWorks colleagues are still assigned to FSMA. Meanwhile, David’s expertise was required at UCB Pharma. ‘UCB Pharma was looking for a consultant to further develop the time series models they had created to be able to predict the sales figures and profit margins of drugs, among other things. Time series are quite complex, but luckily, I had some experience in this area. We managed to make the models smarter so that they provide ever more valuable insights. Thanks to the models, the business analysts at UCB can base their predictions on data – supplemented with their expertise, experience and a bit of gut instinct.’
We managed to make the time series smarter so that they provide ever more valuable insights. The business analysts at UCB can now base their predictions on data, supplemented with their expertise and experience.
Data-driven approach for VDAB
David has also been working on an AI based tool for VDAB, the public employment service of Flanders. ‘VDAB is extremely committed to using data science.’ For instance, VDAB is developing tools based on data and AI to support mediators during the mediation process. A second project David is collaborating on is still in its initial stages: VDAB wants to go through vacancies to map the job competencies of tomorrow. The insights gained will help them prepare applicants for the job market of the future.
Beyond proof of concept
‘Our assignments are very diverse. VisionWorks has an impressive record of achievements, but was operating a little too much under the radar,’ David feels. ‘To position ourselves more strongly in the market, we have recently brought in experts in specific fields, such as Computer Vision, Reinforcement Learning and NLP.’ Why should companies call in Ordina to help them make the most of their data? ‘We do more than just develop a prototype: we also roll out the tool, integrate everything into the existing systems and provide support. And: we are very good at listening to our customer. I didn’t study sociology for nothing,’ David laughs.