🤯 Most data scientists don't know that we can carry out unbiased causal inference under hidden confounding.

(Sometimes)

🧵 (1/n)

#causality #causalinference #python #machinelearning #estimation #causaltwitter Imagine that you work on a causal model.

You're almost sure that there are unobserved variables influencing both - your treatment, and your outcome.

🧵 (2/n)

😯 The greatest fear that drives people away from causality is not about complexity

It's about something much simpler

🧵 (1/n)

#causality #causalinference #python #machinelearning #datascience #casualtwitter If you ask people who have some knowledge about causality, but they never worked in the field, you can often hear that using causal methods in the real world is "too risky".

🧵 (2/n)

😳 Correlation does not imply causation, but does causation imply correlation?
Here's a secret 👇🏼👇🏼👇🏼

🧵

#causality #causalinference #python #machinelearning #datascience #statistics The figure above presents a face-like pattern produced by a data-generating process with very little randomness (strong causation x -> y), but Pearson correlation is essentially zero (no correlation).

(2/n)