A revenue number that misses target raises questions right away. Which region is behind? Which segment? Which accounts? On many teams the answers live in exported spreadsheets or static report decks. They go stale as soon as the data changes, and every follow-up question means another export or another request to an analyst.

Snowflake Dashboards (public preview) give the team one shared source of truth that they can also talk to. Analysts build a dashboard by describing it to Cortex Code (CoCo), then share it with the right roles. Anyone with access can open it in Snowflake CoWork and ask it questions: summarize what it shows, explain what drove a change, or narrow it to one segment. The answers come from the same governed data the charts use. That data never leaves Snowflake, and your existing row access and masking policies still apply.

Under the hood, a dashboard is a single .dash file in Snowsight Workspaces. Each tile runs live SQL against tables or a semantic view in your account. Filters, chart actions, and hierarchy drilldowns take you from a headline number to the rows behind it without leaving the page.

The dashboard we're building is called Revenue Pulse. It runs on synthetic data for a fictional B2B software company: 54 accounts, 3 product lines, and 24 months of revenue through September 30, 2026. The data has a story built in. Q3 2026 revenue is $1.09M behind target, and most of the gap sits in EMEA Enterprise.

You'll set up the data and deploy a ready-made dashboard in your own account, so you get value from it right away:

1. Setup: download three files, point them at a schema you already have, and run one SQL file that loads a small synthetic dataset and a semantic view in about 10 seconds.
2. Deploy the dashboard: upload the dashboard file to Workspaces and open it.
3. Investigate the shortfall: brush a date range on the trend, click a region to filter every chart, and drill from region to segment to account.
4. Ask why in CoWork: publish the dashboard to Snowflake CoWork and ask why the numbers moved.
5. Clean up: run a cleanup script that removes everything.

The diagram shows how the pieces fit together. setup.sql creates three tables and a semantic view on top of them. The dashboard queries the semantic view, and publishing it creates a CoWork artifact that a viewer role can open. cleanup.sql removes the tables and semantic view.

The result is a four-tab Revenue Pulse dashboard running on live queries. The tabs are Overview, Hierarchical Chart, Hierarchical Table, and Data Source.

Everything runs in Snowsight. You don't need Python, a CLI, or a Git integration.

The three files you need are in the getting-started/ folder of the companion repository:

On each file's GitHub page, select Download raw file.

TEMP and CNANTASENAMAT are the author's own database and schema, used here as placeholders. Replace them with a database and schema of your choice that you can create objects in. Do this before uploading the files.

1. In setup.sql and cleanup.sql, replace TEMP and CNANTASENAMAT in the two USE lines near the top. Each line has a comment marking what to replace:

   USE DATABASE TEMP;   -- Replace TEMP with your own database
   USE SCHEMA CNANTASENAMAT;   -- Replace CNANTASENAMAT with your own schema in that database
   

   For example, USE DATABASE ANALYTICS; and USE SCHEMA MY_SANDBOX;. Everything after these lines uses names relative to that schema.

2. In revenue_pulse.dash, replace every TEMP.CNANTASENAMAT with your own DATABASE.SCHEMA, for example ANALYTICS.MY_SANDBOX. The dashboard's queries use fully qualified names, so they work for anyone you share it with.

1. In Snowsight, go to Projects » Workspaces.
2. In My Workspace, select + Add New » Upload Files and upload setup.sql.
3. Open setup.sql. In the editor, pick your role and a warehouse.
4. Select Run All.

The script finishes in about 10 seconds. Its last statement returns one row:

If you see those numbers, the data is ready. The script is safe to run again: it replaces the tables and semantic view each time, and keeps any grants you've added to them.

1. Select + Add New » Upload Files again and upload revenue_pulse.dash.
2. Open revenue_pulse.dash. Snowsight renders it as a dashboard rather than as text.
3. Select Run all if the tiles don't load on their own.

The KPI row shows the Q3 2026 headline: net revenue of $6.9M, up 0.8% year over year, 86.3% target attainment, a variance of about −$1.09M (the card rounds it to −$1M), and 71% gross margin.

Chart actions are interactions an author adds to a chart. When a viewer clicks or drags on that chart (the source tile), other tiles (the targets) respond. A filter action re-runs their queries for just the selection. A highlight action keeps all the data and dims what doesn't match. Revenue Pulse already has filter actions wired up, so you can use them right away.

The dashboard isn't a static report: every chart doubles as a filter. The animation below scrolls from the KPI cards through the trend, regional, heatmap, driver, and account charts. It then clicks a region on the variance-by-region chart, and the rest of the dashboard narrows to that region. That click is one of several chart actions. In the next sections you'll brush a date range on the trend, click a region or heatmap cell to filter every chart, and drill down from region to segment to account.

Try each of these on the Overview tab, and clear the selection before the next one.

Drag across a range of months on Monthly Net Revenue vs Target. The chart zooms to those months, and the KPI cards, regional bars, and heatmap recalculate for the range. Clear the Filtering chip on the tile to undo it.

Click the EMEA bar on Variance to Target by Region. The other tiles filter to EMEA, where attainment drops to about 70%. Clear Drill: region path at the top to reset.

On Target Attainment, the darkest cell is EMEA Enterprise at 57.9%. Click it to filter the rest of the dashboard to that region and segment. A Filtering EMEA, Enterprise chip appears above the heatmap, and every other chart updates to that slice. The scorecards drop to about $1.1M net revenue, 57.9% attainment, and a variance of about −$779K, and the variance bars switch from regions to the six EMEA Enterprise accounts. Select the × on the chip to clear it.

Open the Hierarchical Chart tab. Click EMEA, then Enterprise, to reach its six accounts. The largest miss is Demo EMEA Enterprise 05, at about −$349K. Use the level trail to return to the top.

The Hierarchical Table tab shows the same region › segment › account path as an expandable grid.

Set Region = EMEA and Segment = Enterprise from the filter bar. These filters apply across every tab, which makes them the quickest way to compare one slice against another.

The dashboard has shown you where the gap is: EMEA Enterprise, and three of its accounts in particular. The next question is why. Publish the dashboard to Snowflake CoWork and you can ask that question in plain language, right next to the charts. Anyone you share it with can do the same.

1. With revenue_pulse.dash open, select Share.
2. Add a role with Can view access, then publish. Your own role works; add a teammate's role too if you want them to see it.
3. Open Artifacts in CoWork and select Revenue Pulse.

Viewers run the dashboard with their own privileges, so their role needs SELECT on the three tables and the semantic view, plus a warehouse. Publishing needs the MANAGE ACCESS ON DASHBOARD ARTIFACTS privilege, which is granted to PUBLIC by default.

In the box below the dashboard, ask:

CoWork runs queries against the same data and answers in a side panel, with a table of the drivers behind the miss. In the run below, it found EMEA Enterprise finished Q3 2026 about $779K below target, at 57.9% attainment. Renewal delays explained about 44% of the shortfall, a contract downsize about 33%, and discounts and service credits about 23%. Your answer may be worded differently, but the numbers come from the same data.

Try a follow-up of your own, such as which accounts to call first, or how EMEA Enterprise compares with the same quarter last year.

When you're done:

1. Remove the dashboard's roles in Share and delete the Revenue Pulse artifact in CoWork.
2. Upload cleanup.sql, open it, and select Run All with the same role you used for setup.sql. It drops the four objects the setup created and leaves your schema in place:

3. Delete setup.sql, revenue_pulse.dash, and cleanup.sql from your workspace.

You loaded a dataset, deployed a dashboard file, traced a $1.09M revenue gap to EMEA Enterprise, and asked CoWork why it happened. The .dash file you opened holds every tile, filter, chart action, and SQL query, so it's easy to version, share, and reuse.