Country or region and language

    Choose your country or region and your language.

    Suggested for you

      All countries and regions

        Choose your character

        It shows in the top bar and on your profile. Change it any time.

        Looker dashboard tutorial with a checked example

        A dashboard starts with a query you trust. Define its grain and measures, save the result as a tile, then check that filters and access behave as intended.

        Checked 14 November 2026

        A dashboard brief

        A fictional equipment retailer needs a weekly review of completed orders. Its operations manager wants revenue by sales channel, an order count and a list of regions with falling sales. This brief sets the model’s grain: one order, not one order line or one customer.

        Before drawing charts, decide whether refunded and canceled orders belong in each measure. Write the date field and the comparison period beside that decision. A dashboard can look consistent while its tiles answer different questions.

        A checked Explore query

        In an orders Explore, select the channel dimension and the order-revenue measure. Apply the completed-order filter, run the query and inspect the table before choosing a bar visualization. Sort by revenue so the chart has an order the reader can explain. Check the order count separately from the revenue total.

        The Looker and LookML course starts with these Explore controls. Its dashboard lesson then saves queries into a practice dashboard. The simulator uses original fixture data and does not create a dashboard in Google’s service.

        A join that doubles revenue

        Suppose order A has revenue of 80 and two items; order B has revenue of 40 and one item. The original total is 120. A raw join from orders to items repeats order A’s revenue and gives 200 if you sum the joined rows. Those figures are an illustrative example, not a benchmark or a real store’s sales.

        Define a unique, non-null order primary key and declare the actual relationship. Looker uses model metadata and supported symmetric aggregates to handle measures across fanout. A distinct customer count will not repair an order-revenue sum. Practice join fanout before treating a joined dashboard total as evidence.

        Tiles and filter scope

        A query tile has its own query. A Look-linked tile depends on the saved Look it references. Decide which relationship you want before reusing the same result across dashboards. Changing a shared Look can affect several linked tiles.

        Add a dashboard filter only after deciding which field it should control in every tile. Apply a narrow period or region and confirm that all intended tiles change. A tile can keep its own filter or use a different date field. Our dashboard filter practice checks the scope, rather than treating a visible filter as proof that it works.

        Access and delivery

        Folder access, model permissions and data restrictions serve different purposes. A hidden field is a presentation choice and does not establish row-level security. A dashboard filtered to one region is also insufficient if the viewer can remove that filter. The underlying model must enforce the intended restriction where required.

        Test with a viewer who should see a small subset of rows. Check both the visible chart and the rows available through exploration or download. Scheduling, external delivery and production permissions need a real Looker instance; the practice dashboard does not send reports.

        For the lighter report-builder workflow now called Data Studio, use the Looker Studio tutorial. The comparison guide explains where the two products differ.

        Sources

        Your Work Gym account

        Save your progress and pick up on another device. Your browser's progress comes with you.

        or

        You can keep practicing without an account.

        Help and feedback

        Ask a question, tell us what is not working, or share an idea. We reply by email.

        What is it about?

        We use it only to reply to you.