Notes on Data and Learning
Federica Gazzelloni’s newsletter on data, statistics, actuarial science, reproducibility, AI, and learning through practice.
My newsletter brings together working reflections on data, statistics, actuarial science, reproducibility, AI, and the ways we learn to work with them.
I explore the questions behind analytical work: how we frame a problem, what our models measure, and what changes when we put them into practice.
Selected issues
When Performance Is Not Learning
N.30 · 2 October 2026
How evaluation scores depend on the questions, data, metrics, and conditions used to produce them.
When the Model Changes the Data
N.29 · 18 September 2026
How decisions supported by a model shape future observations and create feedback loops in analytical work.
Before the Pipeline, There Is a Question
N.28 · 13 September 2026
Why defining an analytical question matters when AI helps prepare data, build workflows, and interpret results.