Jamilla Cooiman, owner of Causal Academy, provides in-company training for data teams that want to build stronger practical skills in causal inference.
The trainings are designed for data professionals who want to develop or deepen their skills in causal inference. This includes data scientists, data analysts, product analysts, and related analytical roles.
The training is especially relevant for teams working on questions related to impact evaluation, targeting and personalization, pricing, marketing effectiveness, or other decision problems where causal reasoning matters.
Participants are expected to have experience working with data and basic statistical concepts. No prior knowledge of causal inference is required. When teams already have experience in the field, the level and focus of the training can be adjusted accordingly.
The content of each training can be adapted to organizational needs. Topics that can be covered include:
Fundamentals of causal inference.
Why causal inference matters for decision support, how it differs from predictive modelling, the meaning of causal effect quantities such as ATT, ATE, CATE, and AVSQ, and the core assumptions required for causal analysis.
Preparation stages of causal analysis workflows.
Translating business questions into well-defined causal designs and reasoning about the assumptions underlying the analysis.
Estimation approaches.
From linear regression models to more advanced causal machine learning methods and modern difference-in-differences, applied to both average treatment effects and heterogeneous effects.
Diagnostics and robustness analysis.
Diagnostic tools for assessing core causal assumptions and strategies to address violations, including positivity diagnostics, parallel trends testing and sensitivity analyses.
Below are examples of training tracks that can be delivered depending on the team’s background, goals, and use cases.
A foundational training track focused on building a clear understanding of what causal inference is, why it differs from standard predictive modelling, and what a practical causal workflow looks like in observational settings using the selection-on-observables framework.
Topics can include:
• why causal inference is needed for decision support
• how to translate business questions into well-defined causal designs
• causal targets and their differences
• how the selection-on-observables framework connects unobserved causal targets to quantities that can be estimated from data
• the identification assumptions required for causal effect estimation and how to reason about them using domain knowledge and statistical tools where possible
• estimation using methods such as linear regression, doubly robust estimators and double machine learning
• The use of sensitivity analysis to assess how robust conclusions are to violations of assumptions
A more advanced training track focused on causal analysis in panel-data settings using modern difference-in-differences (DiD) methods.
Topics can include:
• why the classical 2x2 DiD setup is often too restrictive for applied business settings
• how to translate business questions into appropriate DiD designs
• how to apply DiD in staggered adoption settings where treatment timing differs across units
• how to work with treatments that can switch on and off over time
• DiD methods for continuous treatments or changing treatment intensity
• how to use DiD to estimate how causal effects differ across customers, stores, regions, products, or other units
• modern identification assumptions for DiD designs that go beyond the standard binary parallel trends setup
• diagnostics and sensitivity analysis for assessing the plausibility of these assumptions
Trainings can be delivered in a range of formats depending on the team’s goals, background, and availability.
A common format is a series of half-day sessions spread over multiple weeks. This allows participants to build skills progressively and practice with concepts between sessions. Other formats, such as single-session workshops or more intensive full-day trainings, can also be arranged depending on the needs of the team.
Trainings can be delivered online to international teams or on-site within the Netherlands. All sessions are conducted in English (or Dutch if preferred).
The trainings are delivered with theory as the foundation and practice as the primary focus. Methods are discussed in terms of what is ideal from a theoretical perspective, what is feasible in applied environments, and what trade-offs arise in real projects.
Sessions are interactive and include examples, case studies, coding demonstrations, and exercises throughout. Participants retain access to all training materials after completion of the training(s).
Training programmes can be delivered in standardized formats or tailored to organizational needs. Customization can include a focus on specific methodological areas, alignment with internal business use cases, or adjustments in depth depending on the team’s background.
If you would like to learn more or schedule a training, please reach out via email: