Dr. Juan Camilo Orduz

In this notebook we go deeper into the last covariance model presented in the previous blog post Multilevel Elasticities for a Single SKU.

In this notebook I want to experiment with some basic models for price elasticity estimation in the simple context of a simple sku across multiple stores and regions.

In this notebook we present an example of a regression model with time varying coefficients using Gaussian processes.

You will probably be intrigued by the title of this post.

In this notebook we explore regression discontinuity design using generalized linear models (GLMs) and kernel weighting from a bayesian perspective.

This notebook is a continuation of the previous notebook on ATE estimation for binary data with logistic regression based on the sequence of (great!)

In this notebook, I want to reproduce some components of the extensive blog post Causal inference with Bayesian models by Solomon Kurz.

Here I want to share the recording and slides of the webinar Bayesian Methods in Modern Marketing Analytics in collaboration with PyMC Labs.

In this short post, I show how to vectorize an scikit-learn transformer over a numpy array.

This notebook is the continuation of data gathering and data analysis post Open Data: Berlin Kitas.

This notebook presents a simple example of a hierarchical model using NumPyro.

In this notebook we present an example of how to use PyMC to estimate the effect of a treatment in an experiment where there is non-compliance through the use of instrumental variables.

In this notebook we extend the cohort retention model presented in the post Cohort Retention Analysis with BART so that we just model retention and per cohort simultaneously (we recommend reading the referenced post before this one).

In this notebook we study an alternative approach for the cohort analysis problem presented in A Simple Cohort Retention Analysis in PyMC.

In this notebook we present a simple approach to study cohort retention analysis through a simulated data set.

Introduction In this notebook I describe and present an implementation of the time based regression (TBR) approach to marketing campaign analysis in the context of geo experimentation presented in the paper Estimating Ad Effectiveness using Geo Experiments

In October 2022 I had the opportunity to give a talk at the Helsinki Data Science Meetup hosted by Wolt.

In this notebook we want to describe how to port a couple of classification examples from scikit-learn’s documentation (classifier comparison) to PyMC.

Synthetic control can be considered “the most important innovation in the policy evaluation literature in the last few years” (see The State of Applied Econometrics: Causality and Policy Evaluation by Susan Athey and Guido W.

In this notebook I want to reproduce in PyMC the methodology described in the amazing blog post Modeling Short Time Series with Prior Knowledge by Tim Radtke to forecast short time series using bayesian transfer learning 🚀.

In this notebook we want to illustrate how to use PyMC to fit a time-varying coefficient regression model.

On Friday 13th of May 2022 I was invited to join Alexey Grigorev in an event organised by DataTalks.Club to talk about Machine Learning in Marketing.

In this notebook we present a simple example of uplift modeling estimation via meta-models using causalml and scikit-uplift.

In this notebook we describe how to fit Fader’s and Hardie’s gamma-gamma model presented in the paper “RFM and CLV: Using Iso-value Curves for Customer Base Analysis” and the note “The Gamma-Gamma Model of Monetary Value”.

In this notebook we show how to port the BG/NBD model from the the lifetimes (developed mainly by Cameron Davidson-Pilon) package to pymc.

In this notebook we present a concrete example of estimating the media effects via bayesian methods, following the strategy outlined in Google’s paper Jin, Yuxue, et al.

In this notebook we want to experiment to the new KTR model included in the new orbit’s release (1.1).

In this notebook I want to deep-dive into the idea of wrapping a statsmodels UnobservedComponents model as a bayesian model with PyMC described in the (great!)

In this notebook we provide a python implementation of the lab from the Survival Analysis - Chapter 11 of the second edition of the book An Introduction to Statistical Learning (Second Edition).

In this notebook we want to test various ways of getting a better understanding on how non-trivial machine learning models generate predictions and how features interact with each other.

In this notebook I want to describe how to create features inside scikit-learn pipelines using patsy-like formulas.

In this notebook I explore the glm module of PyMC3.

In this notebook we translate the forecasting models developed for the post on Gaussian Processes for Time Series Forecasting with Scikit-Learn to the probabilistic Bayesian framework PyMC3.

In this post we show how to fit a simple linear regression model using TensorFlow Probability by replicating the first example on the getting started guide for PyMC3.

In this notebook I want to explore some data I found on the Berlin Open Data portal daten.berlin.de.

In this post we want to revisit a simple bayesian inference example worked out in this blog post.

In this notebook I want to collect some useful visualizations which can help model development and model evaluation in the context of regression analysis.

In this notebook we want to go take a look into the distributions module of TensorFlow probability.

We describe how to generate a basic disease spread simulation.

In this post I want to explore the ideas behind spectral clustering.

In this notebook we run Monte Carlo simulations to estimate the volume of the \(d\)-ball \[ B^{d}:=\{x \in \mathbb{R}^d : ||x|| \leq 1\}. \] There are many ways to obtain a closed formula for this volume , see for example this Wikipedia article.

In this notebook we are present an initial exploration of the Prophet package by Facebook.

In this notebook we explore the Structural Time Series (STS) Module of TensorFlow Probability.

We describe how to set up a toy-model repository to train and dockerize a machine learning model with data store on aws s3.

We give some LaTex code to create figures of manifolds with boundaries.

In this post I want to show how to use public available (open) data to create geo visualizations in python.

In this post we want to explore the semi-supervided algorithm presented Eldad Haber in the BMS Summer School 2019: Mathematics of Deep Learning, during 19 - 30 August 2019, at the Zuse Institute Berlin.

In this blog post I want to describe the explicit computation of the Laplacian on differential forms on the \(2\)-Torus \(T^2\subset \mathbb{R}^3\).

In this notebook we run some experiments to demonstrate how we can use Gaussian Processes in the context of time series forecasting with scikit-learn.

In this post I present a concrete case study illustrating some techniques to improve model performance in class-imbalanced classification problems.