study
Data Science Winter School explores Machine Learning in one intensive weekend
This January, the BSE Data Science Center held the first edition of a new Data Science Winter School, a two-day course which is part of the continuing education offer at the Barcelona School of Economics.
The participants were a group of 32 researchers, economists, and students from undergrad through PhD level. Among them were 10 current BSE Master’s students, 8 alumni, and two BSE Affiliated Professors. Other academics who attended came from Charles University (Czech Republic), IESE Business School (University of Navarra), and Nottingham University Business School (UK). Professionals who took the course came from Banco de España, the Central Bank of Ireland, and the European Centre for the Development of Vocational Training (Cedefop) in Greece.
A gateway for researchers and analysts in any field
The Winter School was designed and taught by BSE Data Science Center Director Omiros Papaspiliopoulos (ICREA-UPF and BSE). The course was organized into alternating sessions of lectures and hands-on projects for 10 hours on both days. Teaching staff from the Data Science Center provided one-on-one help to students with their projects in-class and were made available for personalized feedback after the course.
“I came into the course hoping to get a good introduction to both the theory and the practice. The course was structured in a way that both got accomplished in a very short time,” said Hannes Mueller (IAE-CSIC and BSE). “The sessions with the professor were followed up by sessions with two tutors who would guide us in the assignments. This helped me to not only get the background and ideas to implement my own programs but to actually write them. This combination was worth pure gold.”
Designed as a gateway for researchers and analysts in any field, the course covered in equal measure programming and theory. Students were introduced to using Python for Machine Learning, while also being taught the theoretical foundations, statistical intuitions, and appropriate use cases of common techniques such as cross-validation, regularization, and decision trees.
“The Machine Learning with Python course at the BSE is a great gateway to the empowering world of programming,” said Ece Yagman, a researcher at Universitat Pompeu Fabra and alum of the BSE Master’s in Economics. “Omiros is exceptionally talented at teaching the material in a "user-friendly" way that makes it possible to grasp the inner workings of Machine Learning and its implementations in one intensive weekend. I would highly recommend this course to researchers in any field who would like to push their work to new limits!”
Left: Participants worked through exercises learning the basics of data analysis in Python and Pandas. Right: Interactive programming notebooks, given to all participants, with all the code used to create the examples from the lectures.
Course participant Liudmila Alekseeva has worked as a Senior Audit Consultant for PwC and as a strategy specialist for Uralsib Bank in Russia, and she is currently a PhD candidate in Finance at IESE Business School. She thought the course gave an excellent overview of the theory and methods in Data Science using Python.
"I enjoyed the BSE Data Science Winter School, it met all of my expectations. Intensive lectures, practical examples, and three individual projects that we had to complete during the course made the learning process very effective," she said. "I also appreciate that the professor not only gave us tools, but also explained the theory underlying them and the potential problems related to their implementation. After this weekend I did not become a professional data scientist in Python at once, but the course gave me a fast start and a great motivation to continue developing in this field."
Another participant who said he thoroughly enjoyed the course was Konstantinos Pouliakas, an Expert on Skills and Labour Markets at the European Centre for the Development of Vocational Training (CEDEFOP) based in Greece. "As a researcher/professional of an EU agency, I was in need of a fast and succinct introduction to the principles of machine learning/Data Science. The course was very rigorous and covered targeted material and Python programs that may act as a springboard for more reading and learning on the topic. One definitely needs more time to digest and practice the material taught, but the course did a good job highlighting to potential Data Scientists that one needs rigorous theoretical understanding of the behavioral models, as opposed to simply programming some lines of code. I would highly recommend the course to professionals who, like myself, do not have the luxury of time due to family or other constraints to attend lengthy courses and lectures."