InTowards Data SciencebyMatteo CourthoudUnderstanding Synthetic Control MethodsA detailed guide to one of the most popular causal inference techniques in the industryJul 30, 202216Jul 30, 202216
YukioMulticollinearity is not correlationIntroduction: The problem with tutorialsJun 12, 202210Jun 12, 202210
InTowards Data SciencebyDavis TreybigThe Experimentation GapHow statistical decision making has evolved, yet most companies haven’t kept upFeb 2, 202216Feb 2, 202216
InTowards Data SciencebyMatteo CourthoudUnderstanding CUPEDAn in-depth guide to the state-of-the-art variance reduction technique for A/B testsJun 29, 20225Jun 29, 20225
InSwiggy Bytes — Tech BlogbyBhavi ChawlaMulti-objective ranking using Constrained Optimization in GBTsCo-authored with Shubha Shedthikere and Jairaj SathyanarayanaJul 5, 2021Jul 5, 2021
InTowards Data SciencebyRaimi KarimIntuitions on L1 and L2 RegularisationHow can they prevent overfitting?Dec 26, 201832Dec 26, 201832
InMeta Analytics BlogbyAnalytics at MetaEstimating the long-run value we give to our users through experiment meta-analysisWe are part of the Facebook Feed Analytics team and we wanted to share methodologies that we’ve developed at Meta around how we create…Feb 16, 20226Feb 16, 20226
InThe Glovo Tech BlogbyOscar ClavijoCluster-randomized experiments at GlovoWe discuss the use of cluster-randomized experiments, when they are used and why it is important to analyze their results in a correct way.Jan 20, 2022Jan 20, 2022
InNetflix TechBlogbyNetflix Technology BlogExperimentation is a major focus of Data Science across NetflixMartin Tingley with Wenjing Zheng, Simon Ejdemyr, Stephanie Lane, Colin McFarland, Andy Rhines, Sophia Liu, Mihir Tendulkar, Kevin…Jan 11, 20223Jan 11, 20223
Edoardo ContiOffline Policy Evaluation: Run fewer, better A/B testsHow offline policy evaluation works, examples on how to use it, and lessons learned from building OPE at FacebookJun 10, 20211Jun 10, 20211
InMercado Libre TechbyPedro ZenoneCausal Inference — Estimating Long-term EngagementMeasuring the impact our models generate in the business can be as complex as training and deploying themJul 8, 20211Jul 8, 20211
InNetflix TechBlogbyNetflix Technology BlogInterpreting A/B test results: false positives and statistical significanceMartin Tingley with Wenjing Zheng, Simon Ejdemyr, Stephanie Lane, and Colin McFarlandOct 7, 202111Oct 7, 202111
