Launch Of Modeling Mindsets Book 🐙

How having more than one modeling mindset has helped me get unstuck on some modeling projects.

Today’s post is a bit shorter — I injured my knee and I launched my new book, Modeling Mindsets - The Many Cultures Of Modeling Data, today.

So I write to you with both joy and pain.

---

Modeling Mindsets is my first book since I decided to become a full-time writer.

And it's the book I wish I had read a few years ago to save myself a lot of time.

Getting Stuck In Modeling Projects

Earlier in my career, I sometimes got stuck in modeling projects. I don’t mean coding mistakes, even though a bug can cost you lots of time as well.

No, by “stuck” I mean that my approach to modeling didn’t even contain the right language to solve the project.

And the worst about it: At first I didn't even realize that it was because I applied the wrong modeling mindset. I studied statistics, which gave me an almost pure frequentist statistical modeling mindset. Back then, I had a narrow lens on modeling.

Here are two particular examples where I got stuck:

• We studied the effect of a drug on a disease outcome. The drug didn’t seem to help, judged by the p-value of the coefficient. Maybe we used the wrong model and or the wrong distribution assumptions?

  As it turned out, the solution was using causal inference.

• I tried my luck in my first Kaggle competition — with my frequentist statistics mindset from university. So instead of random forests and cross-validation, I carefully crafted a generalized additive model using in-sample evaluation (I think R-squared), which turned out to produce bad predictions.

  The solution was to adopt a supervised learning approach.

My first reaction was always to double down on my current mindset. You don’t want to know how many different GAMs I uploaded in my first Kaggle competition 🤣.

But as time went on, I realized that I had to approach the problem differently. If I had had at least an intuition about causal inference and supervised learning, I could have saved myself a lot of time back then.

Understand Many Mindsets

My idea with the book Modeling Mindsets was to take a bird's eye view of approaches from Bayesian inference to unsupervised learning.

It can take years to develop a feel for a modeling mindset, as introductions are often loaded with math and theory.

Modeling Mindsets shines a light on core beliefs, strengths, and limitations of modeling approaches, all while being

• short,

• math-free,

• intuition-first.

It's the book I wish I had read some years ago. And that, I guess, is always a good motivation to write a book.

You can get the book directly from Leanpub or sign up here to get a 30% discount until Friday:

If you think a colleague or a friend could also benefit from Modeling Mindsets, I’d appreciate it if you share the link!