09-20, 11:05–11:40 (Europe/Amsterdam), Escher
Applying the tools and techniques from causal effect estimation in a practical setting can be challenging. Randomized experiments, as the golden standard method for effect estimation, are often not practical. Alternative solutions use observational (non-experimental) data, while they introduce their own challenges, which will be addressed in this talk. These challenges are often not elaborately discussed in text books and can be summarized as follows: 1) samples from treatment group are not available all at once and could become available throughout time (online data stream), 2) appropriate control groups are not immediately available for comparison with the treatment group, and 3) the outcomes of choice in causal effect estimation should be in line with the business questions and accepted KPI’s in the domain.
This self-contained talk targets generic data scientists by presenting the theory of causal effect estimation in a simplified and visual (little-math) fashion. In addition, technical & business requirements, lessons learned from e-commerce & banking, and results are shared when it comes to applying causal effect estimation in practice.
Randomized experiments are the golden standard method for causal effect estimation; however, often they are not the preferred solution due to being impractical, expensive, or even unethical. Therefore, tools and techniques such as potential outcome and structural causal models are introduced to estimate the effect of a certain action on an outcome of interest using observational (non-experimental) data. The theory behind these techniques are presented visually in less-technical tone, suitable for general audience.
The existing causal inference text books and tutorials often take a clean data set with segregated treatment and control group for granted. In practice, however, appropriate control groups are not immediately available and need to be properly defined. This talk aims at providing guidelines in defining appropriate control group for a given treatment group. In addition, guidelines for carrying out continuous causal effect estimation with an online stream of observational data from treated samples are proposed.
The talk elaborates on the e-commerce & banking business questions that causal effect estimation is key in answering them. In addition, it introduces the set of business KPI’s as the outcome where business is interested to know the impact of their marketing efforts with e-mail and phone-call campaigns.
Talk Outline & Structure:
- Introduction [2 min]
- Effect Estimation in Experimental vs Observational setting [10 min]
- Introduction to Randomized Experiment
- Introduction to Potential Outcome & Structural Causal Models
- Effect Estimation in Practice [10 min]
- Problem Definition & Business Objectives
- Defining Control Groups
- Effect Estimation Results
- Summary & Conclusion [3 min]
- Q&A [5 min]
Danial is a data scientist & analytics translator with a PhD in applied mathematics (systems & control). In his career, he has experienced different sectors, i.e. manufacturing, cybersecurity, healthcare, and finance. Danial interest lies in the area of predictive & causal modelling theory and its application. In his current adventure at ABN AMRO bank, he contributes to e-commerce data-driven solutions to improves clients experience, engagement, and satisfaction.