Esse mês o tema é Inferência Causal! E temos convidados internacionais, referências na área: Nick Huntington-Klein vai falar sobre como identificar uma relação de causalidade, e como não confundir causalidade com correlação; e Sean Taylor vai nos trazer a relação da teoria com experimentos para estimar os efeitos causais. O host da noite será Matheus Facure, referência em inferência causal aqui no Nubank!

*** ATENÇÃO: O evento será em inglês, com tradução simultânea para português. Será transmitido online via Zoom, às 19h (horário de Brasília). Para facilitar, clique em "Adicionar a Agenda" no menu lateral ou acima. Atenção que pode haver bug com horário de verão. Por favor registre-se para o evento no link acima!
---------

This month the main topic is Causal Inference! We have 2 international guests, well-known experts in the field: Nick Huntington-Klein will talk about causal identification, making sure we don't confuse correlation for causation; and Sean Taylor will discuss the interplay between theory and experimentation for estimating causal effects. The host for the night will be Matheus Facure, our Causal inference specialist here at Nubank!

Event will be in English, broadcasted via Zoom at 7pm (Brasilia Time). Please register for the event in the link above.
---------

Agenda:
- “Causal Identification” - Nick Huntington-Klein, assistant professor of economics at Seattle University
- “The Relationship between Experimentation and Causal Inference” - Sean Taylor, Data Science Manager at Lyft

Below you can find mode details on each talk and speaker.

Title: Causal Identification

Description: Predictive and classification modeling in data science is, by and large, data-driven. These models are highly effective at answering questions of "what." However, answering questions of "why," which includes questions concerning causal relationships, requires a theory-driven approach. Identification is the process of ensuring that a given statistical analysis actually answers the question of interest, for example making sure we don't confuse correlation for causation. This talk will discuss the process of identification, and pragmatic details of taking a theory-driven approach to analysis in domains without strong existing theoretical underpinnings, or where there is meaningful skepticism of theory.

Speaker: Nick Huntington-Klein is an assistant professor of economics at Seattle University. He is well-known for his work on econometrics and causal inference. This includes developing methods to disentangle cause and effect, and helping to apply them in many different academic and business real-world settings. He is most well-known for communicating how these causal inference methods work to a broader audience of academics, students, data scientists, and businesspeople. He is the author of the book The Effect, as well as a number of other widely-circulated materials on causal inference.

===================================================================
Title: The Relationship between Experimentation and Causal Inference

Description: Two quite different approaches are commonly taken to estimating causal effects: one involving careful theorizing and reasoning about identification and another which relies on randomized interventions and can often ignore theory completely. In this talk I discuss the interplay between these two approaches, what each perspective can learn from the other, and how we can leverage both in order to reliably make improvements to business and social outcomes.

Speaker: Sean Taylor serves as Data Science Manager at Lyft, where he is focused on building advanced analytics tools for forecasting, experimentation, and causal inference. From 2012-2019, Sean was a Research Scientist and Manager at Facebook, on the company's Core Data Science Team. Sean holds a Ph.D. from NYU’s Stern School of Business, with a concentration in Information Systems.

======================================================
About the host
Matheus Facure is an Economist made Data Scientist. He has worked with Credit Card Collections for most of his career, but now he leads Data Science efforts in Credit Card Underwriting. Facure spends his days (at least the good ones) running regressions and breaking down the cause and effect of complex data systems.