"All models are wrong, but some are useful". This quote attributed to different scientists shows the vision of statisticians who know that mathematical models may be a reasonable (but never completely accurate) representation of the structure underlying a data generation process. I will first discuss the importance in applied statistics of choosing a model with an appropriate structure to explore, analyze and extract value from data. As researchers in Statistics, my colleagues and I build such mathematical models that are used to capture the structure behind data. From censored data to copula and from random forests to goodness-of-fit tests, I will present an overview of many contributions that GERAD affiliated researchers in statistics have produced in the recent years and discuss the importance of such findings when creating value from data.
Welcome to everyone!