Modeling in Sigma is a collaborative process that can involve people that traditionally wouldn’t see anything but a dashboard. This collaboration lets business experts lend their knowledge to the model, without the dozens of back and forth emails it generally takes to get the correct information in.
While anyone with the correct permissions can jump in at any point, read on for our recommended approach.
Step 1: Set It Up For Success
Who: BI Team
What: Connect to your data warehouse, set up relationships between tables, and flag the tables to start with.
Sigma is about bringing everyone together in a single platform but it all still starts with the data team. But with us, the set up takes a few minutes or hours instead of weeks or months. Follow the steps below, and everyone in Sigma will be set up for success!
The very first step in starting with Sigma is connecting to your cloud data warehouse. As soon as you connect, Sigma creates a specific type of Dataset called a table based Dataset for all of the tables in your warehouse. Navigate to table based Datasets in Sigma using the existing structure of your data warehouse. What makes table based Datasets different from tables is the additional information you can now add.
Within Sigma, you can set up relationships between tables, add descriptions to table based Datasets, add descriptions to columns, choose which columns are added by default to worksheets created from the table based Datasets and more.
Find Your Core Data
The next step is to identify the tables with the core business data. This is the basic information that forms the backbone of several more specific use cases. This is often customer data, product usage data, demographics, and opportunity data.
You’re looking for the bread and butter tables.
These are the first tables to model and flag, so that business analysts know where to start.
Turn your core tables into Datasets. This will set you up for success in the long run! Datasets allow you to add calculations to your data. Plus, all changes to Datasets propagate forward to all worksheets that are based on the Dataset.
Datasets are also stored in the left rail in the Organization’s folder system. This means it is easy to surface data to the right teams in the exact place they will be looking for it, rather than having people search directly in the connection.
Now that you have the list of the basic tables you are going to model, it’s time to start setting up Links.
Links create a pre-defined join pathway between two data sources. Once a Link is defined, users will easily be able to add columns from the linked data sources to worksheets based on the Dataset or Table. They can use Links to add data using a pre-defined join path and an intuitive UI.
All of the relationships that already exist in your database are pulled into Sigma. If you have foreign keys set up in your warehouse, you should automatically see those relationships in the Links tab in Sigma.
From the Links tab, you can set up additional links within Sigma. Define which data sources to connect to and set up a single or multiple join keys.
All of the work in creating joins is done during modeling, leaving business users to explore the ecosystem of data.
Give context to users with Dataset descriptions and column descriptions. All table descriptions that exist in the database are pulled into Sigma.
The BI team should add additional descriptions that will help guide people unfamiliar with the database. Don’t worry about tracking down information about each column of data, you can give permission to business analysts to add that themselves.
Endorse Your Datasets
Now that you’ve set up your tables and core Datasets, let people know where to start! You can flag tables and Datasets as Endorsed, Warning, or Deprecated. That let’s people know where they should -- and shouldn’t -- start.
When setting a flag, you can add a note as well. Your name and the date are automatically included in the flag. The note can be used to let people know when to use the Endorsed Dataset or to give more depth to a Warning Flag.
Step 2: Set Up Permissions
Who: The BI Team
What: Set up permissions for who can annotate and create from the data in your connected database.
Permissions in Sigma can be granted to teams or individuals. Most of the time, you’ll want to grant permission to teams so you can easily add and remove people from specific sets of permissions. Think about what teams you may want to create, and start there.
Some organizations find setting up teams and permissions based on department to be easiest. Some prefer to set up teams and permissions based on the way people use Sigma, like “Analyst” and “Modeller”.
Teams also give the members a shared, private space to collaborate and share their work on Sigma.
Once you have your teams, it’s time to start granting permissions. In the Connection section of the left-hand navigation menu, you can navigate your connect databases. Go to the Permissions tab on the database object or scope and grant the appropriate privileges.
All permissions in Sigma are additive. If someone is a member of two teams, they will have the permissions granted to both teams.
Everyone with access privileges can build worksheets and Datasets directly on top of the data, and pull in data from the tables via Links.
Step 3: Use What You Got
Who: BI Team
What: Use existing SQL to create Datasets in Sigma
If you just set up your cloud data warehouse and Sigma is the first tool to touch the data, skip to step 4. Otherwise, you probably have chunks of SQL that you’ve been using for analytics. Good news! You can use that work you’ve already done. Take the code you’ve created, and transform it into reusable Datasets that Sigma users can utilize as a base for analysis.
Copy, Paste, Create
Got SQL laying around? You can use it to create reusable Datasets for any user to build on. Just copy and paste your SQL into Sigma’s SQL runner, run it, then click “Save as Dataset”.
Once you’ve turned your SQL code into a Dataset, you can set up Links, add descriptions, and endorse the Dataset the same way you can do with Worksheet based Datasets and tables.
Set up Links between all types of data sources on the same connection. This means you can link your SQL based Datasets to additional relevant information in other Datasets and tables. You can also create Links to your new dataset from your core tables and Datasets.
Let everyone know what data is in the Dataset. If calculations are done within the SQL, we recommend adding a description about the calculation in the column description. This keeps everyone on the same page.
Now that everything is set up, let people know that it’s ready to go!
Step 4: Bring in the Analysts
Who: Business Analysts
What: Add business specific knowledge to Datasets and tables, create worksheets and Datasets with specific business metrics.
Now that you’ve set it up so that people can more easily find their way through the database, it’s time to bring in the business analysts to add in the subject area specific knowledge. Business analysts bring their subject matter expertise to bear when adding in column descriptions indicating exactly what the data is, setting up calculations for metrics and KPIs, and lending additional information about how different data fits together for their use case.
In Sigma, business analysts use worksheets to build out use case analysis, which can then be turned into a Worksheet based Dataset so others can continue to build on their work.
Now that the basic models are set up, invite Business Analysts to add additional context to the Datasets and tables, and start building Datasets for specific use cases.
Worksheet based Datasets are different from tables because you have the ability to build calculations and analysis into the Dataset. This is similar to SQL based Datasets, except rather than writing the SQL you can build the calculations in Sigma.
Worksheet based Datasets are a good way to share key metrics and KPIs, as well as set up a single source of truth for common analytical questions.
Review, Tag, Materialize
Once the business analysts have created the worksheet based Datasets, the data team can review as needed. Worksheet based Datasets can be tagged as Endorsed, so all users know where to start. Large Worksheet based Datasets with complex joins can be put on a materialization schedule to help speed up query time.
Step 5: Benefit from the Work
Who: All Sigma Users
What: Start building and creating Analysis
Modeling with Sigma is an iterative process, that easily continues as people use the system. As new use cases are discovered, new Worksheet based Datasets can be built, materialized and tagged. The best way to find a new use case? Get people in there and using the tool.
We’ve made modeling a collaborative process rather than a top down project.
Set Up Folders
Set up your teams and folders so that users know where to start looking for the data that’s most relevant to them.
Items in Team Workspaces are visible only to members of the team. Items in the Organization Workspace and the folders in the Organization Workspace are visible to everyone in the Organization.
Think About Search
Descriptive titles elevate the correct Worksheets, Dashboards, and Datasets in search. Users can apply filters to easily find exactly what they are looking for. Setting up a naming convention and making sure you are descriptive are the best ways to help people find what they need.