The Major Forms of Data in Data Modeling

Chapter 4

“Begin with the end in mind.” - Stephen Covey.

At a high level, data modeling is meant to be agnostic of the physical implementation details. However, you cannot ignore the physical characteristics of where you’ll store your data, nor the forms of that data you’re working with. Knowing the forms of data allows you to keep the end in mind while you model data. For example, if you’re creating an ML model for an image classifier, your inputs will be images, and your output will be a model that classifies these images.

For decades, most data modeling discussions focused on how data was represented in tables. This was fine when data modeling centered around databases. But the world has changed. Today, many types of data are used in countless ways—structured, semi-structured, unstructured, metadata, and exhaust from ML/AI—and not all neatly fit into the traditional table-centric modeling framework.