Hi Niklas, very nice article and analysis. And also a creative way of thinking in new feature creation. Also, I appreciate that you compared dozens of ML models and their accuracy.

Sometime back, I worked on the same case, and I got accuracy more than 98% in Random Forest. In case if you get a chance to go through the model, I appreciate your inputs.

Link: https://github.com/RutvijBhutaiya/The-Famous-Titanic-Study

Thanks.