How to Set Up Sigma Computing for Analytics

Sigma Computing connects to a cloud data warehouse and presents the data through a spreadsheet-like interface. Setting it up takes four steps: create an account, connect your warehouse, configure team access, and build your first workbook. Most teams are querying live warehouse data within an hour. No SQL knowledge is required.

What Sigma Computing Is

Sigma runs every formula, filter, and chart as a live query against your connected warehouse. There is no data copy, no nightly refresh, and no cube to maintain. When you update a filter or add a calculated column, Sigma sends a query to the warehouse in real time and returns results.

For analysts who know spreadsheets but not SQL, this matters. You can pivot data, join tables, build charts, and share reports using a familiar grid interface. For data teams, it means business users can self-serve without filing a ticket for every new query.

Sigma supports Snowflake, BigQuery, Amazon Redshift, Databricks, PostgreSQL, and MySQL. Most mid-size and enterprise teams using Snowflake find the integration especially responsive because Sigma was built around that warehouse first.

Before You Start

Two things are required before setup.

First, a cloud data warehouse account. If you are evaluating Sigma and do not have one, Snowflake offers a 30-day free trial that covers this use case.

Second, admin or owner-level access to that warehouse. Sigma needs a service account or user with read access to the schemas you plan to query. Without that, the connection test will fail at setup.

Step 1: Create a Sigma Account

Go to sigmacomputing.com and sign up for a free trial. No credit card is required during the trial period. After email verification, you land on the Sigma home screen.

During signup, Sigma asks for a cloud region. Choose the region that matches where your data warehouse is hosted. For a US-based Snowflake account, that is typically US East or US West. Selecting a mismatched region does not prevent connectivity, but it adds latency to every query because data travels farther between Sigma's compute and your warehouse.

Step 2: Connect Your Data Warehouse

From the left sidebar, navigate to Administration, then Connections, then click Create Connection.

Sigma displays a list of supported warehouses. Select yours and fill in the credentials. The required fields vary by warehouse type.

Snowflake requires: account identifier (formatted as orgname-accountname or the legacy account.region format), warehouse name, role, and username plus password. If your Snowflake account uses SSO, OAuth is also supported.

BigQuery requires: project ID and a service account JSON key. The service account needs the BigQuery Data Viewer and BigQuery Job User IAM roles assigned.

Redshift requires: the cluster endpoint, port, database name, and a database user with SELECT access.

After filling in credentials, click Test Connection. If the test passes, click Save. If it fails, the three most common causes are: an incorrect Snowflake account identifier format, outbound connections from Sigma being blocked by a firewall, or the warehouse role lacking the required grants.

Sigma publishes its outbound IP ranges in the documentation under Administration > Connections > IP Allowlist. If your warehouse sits behind a VPC or corporate firewall, add those IPs to the allowlist before retrying the test.

For Snowflake, the minimum grants needed are:

GRANT USAGE ON WAREHOUSE <warehouse_name> TO ROLE <sigma_role>;
GRANT USAGE ON DATABASE <database_name> TO ROLE <sigma_role>;
GRANT USAGE ON SCHEMA <schema_name> TO ROLE <sigma_role>;
GRANT SELECT ON ALL TABLES IN SCHEMA <schema_name> TO ROLE <sigma_role>;

Run these as ACCOUNTADMIN or SYSADMIN before testing the Sigma connection.

Step 3: Set Up Your Team

Open Administration, then People, and invite users by email. Sigma has three built-in permission levels.

Viewer accounts can interact with dashboards, apply filters, and drill into data, but cannot create new workbooks or edit existing ones.

Creator accounts can build workbooks, write formulas, create charts, and publish to shared folders.

Admin accounts have full access including warehouse configuration and user management.

For most small teams, analysts receive Creator access and executives or stakeholders receive Viewer access. This matters for cost control because only Creators and Admins can initiate new warehouse queries. Viewers interact with pre-built dashboards without generating fresh compute charges.

Sigma also supports Teams, which are groups you can assign folder-level permissions to. If your sales team and finance team should see different data, create a Team for each and assign the relevant folders to each group. This simplifies permission management once you have more than ten users.

Step 4: Build Your First Workbook

A workbook is Sigma's term for an analysis document. It can contain multiple pages, and each page can hold tables, charts, pivot views, and KPI cards.

From the home screen, click New, then Workbook. Sigma opens a blank canvas. Click Add New Data Source in the top left. Browse the folder tree to find the connection you created, then select a table.

Start with a table you know well: a sales orders table, customer records, or an events log. Once the table loads, you see columns across the top and rows below.

Common first actions are:

• Add a calculated column: Click the + icon at the end of the column header row and enter a formula. Sigma's formula syntax is close to Excel. DateTrunc("month", [order_date]) truncates a date to the first day of that month, for example.
• Group and aggregate: Click Group on a text or date column to pivot by that field. Then add a numeric column and select Sum, Count, or Average.
• Visualize: Click Visualization at the top right. Sigma recommends a chart type based on the data shape. Switch to a different type using the chart options panel on the left.

When the analysis looks right, click Save. Workbooks default to My Documents. Move a workbook to a shared folder so teammates can access it.

Sharing and Exporting

Sigma workbooks can be shared three ways.

Direct link sharing requires recipients to have a Sigma account. They see the workbook with the same permissions granted to their account type.

Scheduled exports send a PDF or CSV snapshot to email, Slack, or a cloud storage bucket on a fixed schedule. This is useful for weekly summaries that go to stakeholders who prefer not to log in.

Embedded analytics lets you drop a workbook into an internal tool or customer portal. Sigma generates a signed embed URL from a service account. This requires a Creator-level plan or above and configuration in the Administration panel.

If you want to skip warehouse setup entirely and analyze an uploaded file in plain English, VSLZ handles that from a single prompt with no connection configuration required.

Common Setup Issues

Query timed out: The Snowflake virtual warehouse is undersized for the query. Resize the warehouse in the Snowflake console (from XS to S, for example) and consider enabling auto-suspend to avoid idle charges.

Permission denied on schema: The Snowflake role used in the connection lacks USAGE on the schema or SELECT on the table. Run the grant statements from Step 2 and retry.

Connection passes but no tables appear: Sigma shows only schemas and tables the connected role has explicit USAGE and SELECT grants on. Expand the grants or switch to a role with broader permissions.

IP allowlist error: Sigma's outbound IPs are not in your warehouse's network allowlist. Retrieve the current IP list from Sigma's documentation under Administration > Connections > IP Allowlist and add them to your VPC security group or firewall rules.

What Comes Next

Once your first workbook is working, two features are worth exploring early.

Data models are reusable table definitions you build once and reference in any workbook. If five teams all need revenue data with the same currency conversion applied, you build one model with the conversion logic and point every workbook at that model. Changes propagate automatically.

Controls are interactive filters that users can adjust without editing the workbook. You can add a date range picker, a region dropdown, or a free-text search. Controls turn a static report into a self-service tool without any code.

Summary

Sigma Computing setup has four steps: create an account in the right cloud region, connect your warehouse with correct credentials and grants, assign team permission levels, and build a first workbook on a table you know. Most of the friction is in Step 2, specifically permissions and network allowlisting. Once that is resolved, the rest is fast. Teams typically reach their first live dashboard within an hour of starting the trial.