Causal inference
Causal inference is the process of determining or estimating how causes, exposures, or interventions affect outcomes, using data and assumptions about causal relationships.
Causal Inference for Data Science
A causal problem is a problem of determining whether and how one factor causes an outcome.
- How it is used
- Guides decisions about changing treatments, policies, products, and other factors when the goal is to affect outcomes.
- How deep it goes
- Develops workflows, graphs, confounding, ATE, and estimation methods while deferring detailed formulas and later graph tools.
- Examples
- Uses illness treatments, website traffic, arcade revenue, COVID-19 triage, and synthetic game-level data to expose misleading associations.
- What you can do after
- Evaluate/understand whether a causal question needs an experiment or observational analysis, then map confounders and assumptions before estimating effects.
Read how Causal Inference for Data Science explains causal inference →
Causal AI
A method for estimating or reasoning about the causal effects of interventions or exposures from data.
- How it is used
- Turns intervention and explanation questions into data-supported analyses for experiments, observational studies, and decisions.
- How deep it goes
- Develops a query-to-refutation workflow and causal effect estimation, while treating other causal queries as extensions.
- Examples
- Uses an online-game engagement experiment, an e-commerce feature’s effect on sales, and handwritten-digit images.
- What you can do after
- Apply it via DoWhy to estimate causal effects while checking whether assumptions fit the problem.
AI Model Evaluation
Causal analysis estimates the effects of actions or interventions on outcomes while accounting for how observations were generated.
- How it is used
- Estimates whether recommendations, rankings, or product changes caused observed user and business outcomes.
- How deep it goes
- Explains counterfactual reasoning, logging, off-policy estimation, and validation checks, but treats estimates as assumption-dependent rather than perfect replays.
- Examples
- Uses movie recommendations, e-commerce product recommendations, clicks, purchases, and alternative ranking or recommendation policies.
- What you can do after
- Evaluate alternative recommendation or ranking policies before online exposure using logged decisions, outcomes, and policy probabilities.
Also covered in
What to know first
- experimentationExperimentation is the planned investigation of a question or hypothesis by conducting trials or interventions and observing their outcomes.
- artificial intelligenceArtificial intelligence is the field of computing concerned with creating systems that perform tasks associated with intelligence, such as learning, reasoning, perception, and decision-making.
- structural causal modelA structural causal model represents causal relationships among variables through equations that determine each variable from other variables and exogenous factors.
- causationCausation is a relationship in which a change in one event, factor, or condition produces or influences a change in another.
Related concepts
- language modeluses · A language model is a computational model that represents patterns in linguistic sequences to assign probabilities, predict, generate, or evaluate language.
- observational datauses · Observational data are measurements collected without the investigator assigning or controlling the conditions, treatments, or interventions being studied.
- causal identificationhas part · Causal identification is the process of determining whether a causal quantity can be uniquely computed from available observational or interventional information and model assumptions.
- causationimplemented by · Causation is a relationship in which a change in one event, factor, or condition produces or influences a change in another.
- graphuses · A graph is a mathematical structure consisting of vertices and edges that connect pairs of vertices to represent relationships or connections.
- cross-validationcontrasts with · Cross-validation is a model-evaluation technique that repeatedly partitions available data into training and held-out subsets to estimate performance on unseen data.
Questions about causal inference
What is causal inference?
Causal inference is the process of determining or estimating how causes, exposures, or interventions affect outcomes, using data and assumptions about causal relationships.
Which Manning books explain causal inference?
Causal Inference for Data Science, Causal AI and AI Model Evaluation.
What should I know before learning causal inference?
Manning books that teach causal inference usually introduce experimentation, artificial intelligence, structural causal model and causation first.
What is the difference between causal inference and cross-validation?
Causal inference is the process of determining or estimating how causes, exposures, or interventions affect outcomes, using data and assumptions about causal relationships. Cross-validation is a model-evaluation technique that repeatedly partitions available data into training and held-out subsets to estimate performance on unseen data.
What are the parts of causal inference?
Causal inference includes causal identification.
What is causal inference also called?
Causal inference is also called causal analysis.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.


