This week in Wayfair Data Science’s explainer series, Tim O’Connor discusses experimentation in the context of data science. Experiments are crucial to data science, helping to determine which version of a model is to use in future iterations of a system or generating new sources of data unavailable in a company’s historical data. As such, designing and conducting experiments is a core capability for data scientists at Wayfair. In this video, Tim explains the basics behind A/B tests and synthetic controls, and discusses how they are used to tackle various questions at the company.
Tim O’Connor is a data scientist on the pricing team at Wayfair. After growing up on an organic blueberry farm in rural Minnesota, Tim found a love for applied mathematics that led to research in a diverse set of fields, including chaos theory, computer vision, and algorithmic game theory. While eventually focusing on auction theory in grad school at the University of Oxford, he is now minorly obsessed with combining the best of both worlds between causal inference and machine learning. When not designing experiments, he enjoys cultivating blueberries, collecting hot sauces, and perfecting one pot dishes.