Data Modeling and Database Design for Beginners

You inherited a spreadsheet. It began as a list of orders, and it now has a column called notes2, the same customer spelled three different ways, and one tab that only one person is allowed to touch.

This is a beginner course in data modeling and database design, and it teaches the step that comes before anyone creates a table. Data modeling is the work of deciding what your data actually means, then writing that decision down in a form other people can read and argue with, which is the skill nobody teaches you before they hand you the spreadsheet.

It is deliberately code free. No SQL is taught here, and the first lesson says so out loud, so nobody sits through the course waiting for code that never arrives. There is no software to install, no database to set up and no diagramming tool to learn. Notation outlasts tools, and time spent clicking around a drawing editor is time not spent on the idea.

One example runs through all 16 lessons. A small online bookshop that sells books to customers, who place orders. It grows only as much as each lesson needs, and it never turns into a hospital or an airline halfway through. By the last lesson you have watched one design travel from a sentence someone said in a meeting, through entities and attributes, primary keys and foreign keys, an ER diagram in crow's foot notation, into real tables, through first, second and third normal form, and out the other side as a star schema built for reporting.

What you actually get. Sixteen lessons across three sections, under an hour of runtime in total, and eight downloadable files running to roughly 150 pages:

• A prompt pack of 96 prompts, six for every lesson, written for your own data rather than a textbook one

• Three worksheets, one per section, meant to be filled in while you watch

• A workbook with the bookshop example worked from first sentence to finished tables

• A glossary, so a junction table is never also called a bridge table without warning

• A page of errors that are commonly taught as fact, each with the correction

• A further reading list where every source was opened and checked

Where it goes further than most beginner courses. The normalisation lessons are taught the way the original definitions state them: against every candidate key and against non-prime attributes, not the familiar paraphrase about non-key columns and the primary key. There is a whole lesson on functional dependencies first, because the normal forms are written in that vocabulary. And the course does not stop at third normal form. It names what third normal form still lets through, and what BCNF was defined to fix.

What this course does not cover. No SQL queries. No database administration, tuning, indexing or storage internals. No NoSQL, document or graph modelling. No specific tool, so no erwin, no Lucidchart, no Power BI, no dbt. It is also too introductory for a working DBA. Those boundaries are named in the last lesson with a pointer to what to read next.

A note on spelling: data modelling with two Ls is the same subject and the same course.

If you have a spreadsheet you are slightly afraid of and an hour you can give it, enrol and start with the first lesson.