Juan Orduz is a Berlin-based mathematician and data scientist. His work is grounded in a mathematical view of data problems, with particular interests in statistical learning, time series analysis, and Bayesian and geometric methods in data analysis.

That mathematical foundation shows up in how Juan approaches applied problems. In marketing data science, for example, he has worked through attribution, media mix modeling, and uplift modeling: problems where regression, ad-stock effects, saturation curves, and time-series counterfactuals matter far more than the latest model architecture, and where Bayesian methods offer a principled way to handle uncertainty and learning decay.

Juan is also a generous member of the data community. He shares technical writing and open-source projects on GitHub, walking through the methods he uses in practice so that others can apply, critique, and extend them in their own work.