Cookbook: Making Small Dataset Projects Successful (CFD, FEA, …)

Do I have enough data? How do I know? How to quickly fix it if I don't? Let's get some answers in a cookbook/step-by-step type guide.

Step-by-step Guide

There are plenty of good reads out there on the background theory and conceptual fundamentals, but I think there is a lack of succinct instructions in a ‘playbook’ style. Hence, I will try to keep it brief so this remains useful.

This guide aims for a combination of upfront quantitative planning (learning curves, checking sample complexity) and adaptive sampling with validation/UQ to ensure your regression model achieves high (enough, lol) fidelity with the least amount of data, and that you know how much data is enough to trust its predictions.