What is a data model?
1. What is a data model?
Welcome to the course Data Modeling in Sigma!2. Welcome!
In this course, we’ll learn how to build data models in Sigma to set a foundation for scalable, consistent analytics, and enable users to independently act with data.3. Welcome!
I’m your instructor, Ben Harris. I’m an instructional designer and technical writer at Sigma. Along with my collaborators, Joshua Bowman and Andrea Cato, I’ll help you get started with data models in Sigma.4. Oakmark needs your help
In this course, we’ll step into the role of an analytics engineer at Oakmark Bank.5. Oakmark needs your help
Data teams at Oakmark Bank are creating new analyses in Sigma every day. We’ve been asked to create standardized data models, so that teams can make consistent insights, without repeating manual work each time they access our data.6. Defining 'data models'
Let’s start with understanding what a data model is, and why we might use one.7. Defining 'data models'
Data models are a type of document in Sigma, like a workbook. Data models differ from workbooks in two key ways: they do not focus on user interaction, and they can share their data elements with other Sigma documents.8. Defining 'data models'
This means data models are great for curating and governing data across an organization.9. Defining 'data models'
Let’s consider a metaphor. Imagine that Oakmark’s data pipeline is a restaurant. The data warehouse is the kitchen — it stores all the raw ingredients, like customer, loan, and transaction data. The data team is the staff — preparing and delivering the data.10. Defining 'data models'
The data model is the menu. It organizes ingredients into ready-to-serve “dishes.” It defines what’s available, and how it should be presented. That way, the customers (our end-users) can order a complex dish, without having to know how to cook it themselves.11. Defining 'data models'
With data models in place, everyone pulls from the same trusted menu, ensuring the data is accurate, governed, and reusable.12. Defining 'data models'
Most organizations have multiple data models that serve specific use cases, rather than one giant data model. As we move through this course we’ll think carefully about the use case for each data model. This might be in terms of the team using the data, or the types of workbooks, apps, and analyses it will serve.13. Components of a data model
So, what goes into a Sigma data model?14. Components of a data model
Data models begin with a base table. Your base table should have one row for each object or event you’re modeling for users. For example, in a data model for transactions, each row should represent a unique transaction.15. Components of a data model
The base table might include multiple tables from your warehouse. For our base table of transactions, we might join on account data, or use Sigma’s relationships feature, so that users can add columns from other tables when they need them.16. Components of a data model
Then, we might add metrics. Metrics are pre-defined calculations, which provide calculated values that can be used in workbooks right away.17. Components of a data model
Data models may also have parameters to filter data, or be configured with column level security, so data is shown only to users with the correct permissions.18. Components of a data model
There are other advanced features, like materialization schedules, which this course doesn’t cover. However, we believe you’ll have a firm understanding of data models at the end of this course, and will be equipped to explore those features independently in the product and documentation.19. Let's practice!
We’ll cover these components in detail in the next videos. For now, test your knowledge with a few questions.Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.