How did you get started in forecasting?
While studying for my bachelor’s degree in the School of Electrical and Computer Engineering at the National Technical University of Athens (NTUA), I was very fortunate to meet my mentor, Kostas Nikolopoulos, at that time coordinator of the Forecasting and Strategy Unit (FSU) directed by Vassilios Assimakopoulos. Historically, the FSU was employing undergraduate engineering students as junior software developers and assistants in research projects. I joined the FSU in October 2003 and started my first forecasting project just weeks after that. I was fascinated.
In 2005, the FSU collaborated with Spyros Makridakis to develop a state-of-the art business forecasting system called Pythia (named after the Greek Oracle of Delphi). As you can imagine, being in a position to work closely with one of the founders of the field really helped me make my next decision…. I started my PhD in 2007 under the supervision of Vassilios, but working closely with both Spyros and Kostas. The experiences gained in these years were more than enough to close the deal.
How did you get from Greece to the UK?
The completion of my PhD, coupled with the Greek economic crisis, set the scene. A post-doc offer in the Lancaster Centre for Forecasting (LCF) was most timely and meant that my career would progress in the UK. That was certainly a big change in my life, both personally and professionally. Not only was I able to form significant collaborations with colleagues like Robert Fildes and Nikolaos Kourentzes, but I also had the time needed to develop myself as a researcher. Nowadays, I am enjoying my lectureship post at Cardiff University, working closely with Aris Syntetos, Foresight’s Supply Chain Forecasting Editor.
What areas of forecasting interest you?
I am especially interested in research for improving forecasting processes. I strongly believe that such improvements can directly be translated into efficient design of forecasting support systems. So my research focuses on model selection, aggregation in both temporal and cross-sectional levels, as well as effective integration of management judgment. With regards to the latter, I am the cofounder of the Forecasting Society (www.forsoc.net), a platform to promote and disseminate judgmental forecasting research and its applications.
How do you see the future of forecasting support systems?
In my view, future forecasting support systems should be described with three terms: Web-based, open-source, and customizable. Consider the success of the R statistical software and the very powerful packages developed for this platform, an open-source, modern, and customizable Web interface that employs freely available libraries to provide access to state-of-the art forecasting methods.
Do you work with businesses to improve their forecasting practices?
As a member of both the FSU and LCF, I have had the opportunity to collaborate with numerous companies in forecasting-related projects. In many cases, we improved forecasting performance considerably, reducing operational costs. Knowledge dissemination is also taking place at the university level, through research-led teaching and supervision of company-based MSc dissertations. Apart from these, one of the best ways to see your research being directly applied by companies is to make it easily available. To this end, I have coauthored two R packages (MAPA and tsintermittent).
Do you have any free time?
I do. I enjoy watching movies, TV series, walking, and cycling. Apart from that, I had been training in martial arts for more than 15 years, an activity that I may resume in the near future. However, my passion is music; in another life I would be a professional disc jockey!