How to Get Started with Sigma Computing

Sigma Computing lets you query a cloud data warehouse the same way you would use a spreadsheet. You apply filters, create pivot tables, write formulas, and build charts - and Sigma silently converts those actions into optimized SQL running against your live data. No SQL knowledge needed, no data exports, and no lag from cached reports.

What Sigma Is Designed For

Sigma is built for business users who have data in a cloud warehouse but do not want to depend on a data engineer every time they need a new cut of numbers. Instead of exporting a table to Excel, cleaning it up, and building a pivot - which can take hours and go stale within days - Sigma connects directly to Snowflake, BigQuery, Databricks, or Redshift and runs queries on demand.

The platform is warehouse-native. That means your data never leaves the warehouse, permissions are enforced at the source, and reports reflect the current state of the data every time someone opens them. For ops teams and finance teams that track metrics in real time, this is a significant difference from tools that rely on nightly syncs.

Prerequisites

Before you start, you need three things:

A Sigma account. Sigma offers a 14-day free trial with no credit card required. You can sign up at sigmacomputing.com. After the trial, pricing is per-seat with enterprise plans available for larger teams.

A cloud data warehouse. Sigma works with Snowflake, BigQuery, Google AlloyDB, Databricks, Redshift, PostgreSQL, and MySQL. If you do not have a warehouse yet, Snowflake and BigQuery both offer free tiers you can use for testing.

Admin credentials for your warehouse. During setup, Sigma will ask for a service account or a read-only user with permission to query the tables you want to analyze.

Step 1 - Connect Your Data Warehouse

After creating your Sigma account, you land in the Admin panel. Go to Administration > Connections and click Add Connection.

Choose your warehouse type. For Snowflake, you will enter your account identifier (the subdomain in your Snowflake URL), a username, a password or RSA key, the warehouse name, and the default database and schema. For BigQuery, you upload a JSON service account key file.

Sigma recommends creating a dedicated read-only service account for the connection. This limits what Sigma can access and makes it easier to audit later. In Snowflake, you can set this up with a role that has SELECT grants on the schemas you want to expose.

Once the credentials are saved, click Test Connection. Sigma runs a short query to verify access. If the test passes, your warehouse appears in the connection list.

Step 2 - Browse Your Data

With the connection active, go to the main Explore panel and click Browse Data. Sigma shows a catalog of every database, schema, and table your service account can see.

Click any table to open a preview. Sigma loads the first 500 rows and shows column names, data types, and a summary row count. This is the fastest way to verify that the connection is pointing at the right data before you start building anything.

If you manage multiple schemas - say, a production schema and a staging schema - you can pin specific tables to a shared data model later. For now, browsing directly from the connection is enough to get started.

Step 3 - Create Your First Workbook

A Workbook in Sigma is the equivalent of a spreadsheet file. It can contain multiple pages, and each page can hold tables, charts, pivot tables, and input forms.

Click New Workbook from the home screen. You will be prompted to choose a data source. Select the connection you just created and choose a table.

Sigma opens a table view with all columns visible. From here you can:

Apply column filters by clicking the filter icon on any column header. For example, filter a date column to show only the last 90 days.

Group by a column to summarize rows. Drag a column like Region or Product into the group-by zone and Sigma aggregates the rows automatically - the equivalent of a GROUP BY in SQL.

Add calculated columns using Sigma formulas. The formula syntax is similar to Excel. For example, [Revenue] / [Sessions] creates a revenue-per-session column without writing any SQL.

Sort and reorder columns by dragging headers. Sigma does not require you to know which columns exist in advance - the interface is explorative.

As you make changes, Sigma generates SQL in the background and re-runs the query. You can click the SQL icon in the toolbar to see exactly what query is running. This is useful for building SQL literacy over time without being blocked by it.

Step 4 - Build a Chart

With a grouped table on screen, click the plus icon in the page panel and choose Chart. Sigma opens a chart builder on the right side.

Drag columns into the X axis, Y axis, and Color fields. Sigma recommends chart types based on the data - bar for categorical comparisons, line for time series, scatter for correlations.

Click the chart type dropdown to switch between bar, line, area, scatter, and pie. For most operational dashboards, bar charts for snapshots and line charts for trends cover the majority of use cases.

Charts update live when you change a filter on the parent table. This means you can build a filter control once and have every chart on the page respond to it.

Step 5 - Share and Publish

Once your workbook is ready, click the Share button in the top right. You can invite teammates by email, share a view-only link, or publish the workbook as a dashboard with scheduled refresh.

For scheduled delivery, Sigma can email a PDF or image of the dashboard on a daily or weekly schedule. This is useful for stakeholders who want a regular snapshot without logging in.

Sigma also supports embedding dashboards into internal tools using an iframe or the Sigma Embed API. This is commonly used in customer-facing portals where you want to show each customer their own data without building a custom dashboard per account.

What Sigma Does Not Handle

Sigma is strong at exploration, reporting, and dashboarding on clean, structured warehouse data. It is not a data ingestion tool - you still need a pipeline (Fivetran, Airbyte, or a custom ETL) to get your raw data into the warehouse in the first place. It also does not do statistical modeling, machine learning, or freeform analysis on unstructured data.

If your data is still in spreadsheets or uploaded files rather than a warehouse, the analysis loop looks different. VSLZ lets you upload a file directly and get charts, statistics, and plain-English summaries from a single prompt, without needing a warehouse connection or any setup.

Practical Summary

Sigma Computing reduces the friction of working with warehouse data by replacing SQL with a spreadsheet interface most business users already know. The setup process takes under an hour for a team that already has a Snowflake or BigQuery account. The biggest hurdle is usually getting the right service account permissions from whoever manages the warehouse - plan for that conversation before you start.