## ALFRED GALICHON'S

## MASTERCLASSES

Story of math-econ-code

In an age when online training courses are proliferating and are becoming increasingly accessible, Alfred Galichon's 'math+econ+code' masterclass series stands out as an original proposal.

The formula of the masterclass was inspired by cooking lessons: a good cooking class would never put theory first, but would rather work by learning-by-doing, breaking down the problem into sub-problems while emphasizing the importance of preparation. Like these cooking classes, math+econ+code works by placing students in front of a concrete question based on an actual dataset, and takes them on a journey where they craft their own methodology and tools to address the question. Working so by trial and error allows the student to understand the power of the mathematical tools they manipulate, and lets them to acquire the technical proficiency needed for these analyses. In contrast to traditional mathematical classes that usually teach under the "theorem-proof-example" format, 'math+econ+code' reverses the order and places the student in the shoes of researchers.

As its name indicates, 'math-econ-code' combines mathematics, coding and economics, and hence, pertains to data science applied to economics. The objective is to harness data science tools for the purpose of economic analysis. For example, the machine learning libraries TensorFlow and Pytorch can be put to work on vast economic problems.

Over the years, the masterclass series has also allowed to create a community gathering researchers who share the same interests: the courses being different at each session, some students return several years in a row. A seminar series has originated from the masterclass and is attended by members of the community; sometimes the speakers are also alumni of the masterclasses. In addition, the masterclasses have been online-only since the COVID-19 pandemic, which allows students to both better follow and come from all over the world, hence further diversifying the 'math+econ+code' community. As this community has grown, the 'math+econ+code' website has gradually expanded, incorporating notebooks, seminars annoucenments, and a blog.