Assignment: Read Chapters 1–3: The Dawn of Mixed Model Arts, What Data Modeling Is (and Isn’t), Why Data Modeling Matters.
You’ve probably been here, or experienced a version of this problem. A support application can work perfectly for its users and still create a mess downstream. The application team needs to open and update tickets. An analyst needs to compare resolution times. Someone building a recommendation system might want to find similar complaints. They’re working with the same business, but they need different things from its data.
Let’s start here for week 1 of the book club. Chapters 1–3 introduce the different modeling camps and the problem of treating your preferred approach as the answer before you’ve understood the question. The useful part is learning what another approach can show you, then making the representations agree where they need to.
Our running example comes from Chapter 12: an online retailer handles support through email, chat, and phone. You don’t need to read ahead. For this week, imagine the support team wants a simple way to close tickets, while the analytics team needs to know whether the customer’s problem actually went away. A closed ticket may be a useful application state. Whether it answers the business question is something we need to examine.
As you read, notice the decisions that get made even when nobody calls them “data modeling.” Someone chooses what to record, what to leave out, and which definitions to assume. Those choices travel downstream, often in difficult ways toward analysts and data scientists. I’d like us to talk about what a small, useful amount of deliberate modeling looks like before those assumptions become expensive to change.
Questions for discussion
Which modeling camp most influences your first instinct, and what might it miss in this support example?
What must the application and analytics representations agree on for their answers to remain coherent?
Which assumption about a closed ticket would you test with a support agent or customer before building a report?
Optional exercise: Spend ten minutes sketching the support application and a resolution-time report. Write one question each serves and circle one definition they need to share. No schema or code required.
Put a question or sketch in the public comments. If you’re behind on the reading, join with the example above. Remove confidential details from anything you share.
Live sessions and recordings are included for paid Practical Data Modeling subscribers. Check the paid subscriber chat on Substack for the confirmed live day, time, and Google Meet link. Message me there if you need the calendar or Meet link. October 21 is our reading-week target.
See you at book club,
Joe Reis


I like your book club initiative Joe - I'm following in the background from Athens, Greece :)
I waited in the lobby for a while and was not able to join. I hit the Ask to join button too. Am i missing some step?