Why This Matters
An Ask HN post floats the idea of an intuition-first statistics textbook, arguing existing books are poor and pointing to interactive resources like Seeing Theory as a better model. It's a small signal about how technical education is changing: as data literacy becomes core to software and AI work, practitioners are questioning whether static textbooks still serve learners.
Key Takeaways
- The author, trained in Bayesian statistics, argues most existing stats textbooks fail at building intuition.
- Interactive visual resources such as Brown's Seeing Theory are cited as the benchmark for teaching stats well.
- The open question is whether a book is still the right format, or whether effort is better spent on interactive material.
I had a classic Bayesian statistical education throughout my undergraduate and grad years and I’ve come to conclude that the plethora of books offered are pretty bad.
For me personally statistics is intuitive if illustrated properly. A great example is https://seeing-theory.brown.edu/
I was wondering whether it would make sense to turn the intuitions behind statistics into a book. Would anyone read it? Do people still read stats books or would it be mostly a waste of effort? If you wanted to teach/help people understand statistics, what resources would you consider?
GoKawiil