↵ select ↓ ↑ navigate esc close

Week 2 - Recognizing the Same Ticket in Different Forms of Data

Practical Data Modeling ·

Read Chapters 4–6, printed pages 63–124: Beyond Rows and Columns: The Five Forms of Data; Entities, Instances, and Identifiers: Naming What You Model; Attributes: Describing the Entity.

A customer emails support, starts a chat, and eventually calls. You now have a CRM record, a JSON conversation, and a phone transcript. How many support issues do you have? Obviously, the file formats alone won’t answer that question.

Practical Data Modeling is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

This week’s chapters cover data forms, entities, identifiers, and attributes. Chapter 4 separates a data form from the physical format used to store it. Chapters 5 and 6 ask us to be precise about what we’re representing and the properties that describe it. That precision matters when the same business concept appears in several systems, which happens almost everywhere nowadays.

In the Chapter 12 support example, Ticket is a shared entity with a channel attribute, rather than three unrelated concepts called ChatTicket, EmailTicket, and PhoneTicket. Let’s add a hypothetical complication: a customer follows up through another channel, and the systems don’t share a ticket identifier. We need to decide whether these are different representations of the same ticket or separate tickets concerning the same underlying issue. Matching an email address may help, but that’s error-prone and doesn’t settle that decision.

Attributes bring their own questions. A field called date doesn’t tell us whether it means creation, first response, or resolution. A sentiment score extracted from a transcript is also different from something the customer explicitly said. Both can be useful, but future consumers (stakeholders and AI agents) need to know where the value came from and how it was produced. Otherwise, we’re giving them more to guess at.

Questions for discussion

  1. What would make two records representations of the same ticket, and what evidence would you need to decide?
  2. Which ticket attributes can you capture directly, and which would you derive from a transcript?
  3. If two systems disagree about a ticket’s identity or an attribute, what should a downstream consumer be able to see?

Optional exercise: Make a small Ticket definition with five attributes. For each attribute, note its meaning and source. Include one identifier and one timestamp precise enough that someone else could use it without asking you what it means.

Share your sketch or a question in the public comments. You can participate without finishing every chapter. Use made-up records and keep workplace or customer information private.

October 28 is the reading-week anchor. We should be good, but I’m also traveling that day. Check the paid subscriber chat on Substack for the confirmed live day, time, and Google Meet link. Live sessions and recordings are included for paid Practical Data Modeling subscribers. Message me on Substack if you need the calendar or Meet link.

Practical Data Modeling is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.