Below are my recommendations on some statistical resources.

Want to know more about a resource?

Would you like to be accompanied to master the resources presented more quickly?

Your favorite book or software on Bayesian statistics, causality or statistics in general do notcan’t find it?

Contact me!

The ‘Book of Why’ by Judea Pearl is a superb introduction to causality. he explores the history leading to the developments of the new science of causes and effects. And despite his almost ten years, he is already exploring the relationship between causality and artificial intelligence.

The Bible that all good Bayesian should have at home 🙂 More seriously, this book is very explained and illustrated and allows a passage without trauma (or almost) from frequentist statistics to Bayesian statistics, with examples of reproductions of frequentist analyses.

a language to govern them all… I nI will therefore quote thatonly one:

The best Python packages beyond the classics (Pandas, Numpy…):

  • data processing
    • Pyjanitor: To clean the data
    • Polars: a more robust panda for big data
    • Skimpy: a pandas.describe() on steroid
  • datavis
    • Seaborn: very beautiful graphics in very few lines of code
  • Causality
    • Dowhy: for theestimate ofCausal effect
    • Causal-learn: for causal discovery
    • Causalpy: for theBayesian causal inference
    • ECONML: for theCausal AI
  • Bayesian statistics
    • Bambi: Quickly build Bayesian linear mixed models
    • Pymc: the basis for making Bayesian modeling and a little causality
    • Arviz: Essential along side Pymc for visualizations
    • Pymc-Marketing: For mixed marketing models in Bayesian
  • ML, AI and LLMS
    • Scikit-Learn/Pytorch/TensorFlow: For Lartificial intelligence
    • LangChain/Langgraph: For LLMS

All the same, I have totalk about it, because python nIsn’t good everywhere I have toadmit…

For production, jwould use rust when possible, in order toHave good performance and safety

Jasp is a point-and-click software that is very suitable for students or to do quick tests. It includes frequentist and Bayesian methods, which is useful for making comparisons. Additionally, it has multiple plugins to explore various domains or learn. he is free to access

G*Power is the complete software to perform Sample Size estimation. It is free and very powerful

  • https://statmodeling.stat.columbia.edu: Blog of Andrew Gelman, Professor of Statistics and Political Science at the Columbia University, he notably co-created (with Bob Carpenter) Stan which is a very famous bookstore allowing you to do Bayesian modelling in R and Python. This blog focuses on Bayesian statistics and causality

  • https://www.youtube.com/@rmcelreath (English): Professor of anthropology, Richard McElreath gives an open access course that II particularly appreciated on causal modelling using advanced Bayesian techniques. I recommend a binge watching with more than 26 hours of lessons 😉
  • https://www.youtube.com/c/3blue1brown (English): Not just statistics, but mostly math with beautiful 3D animations