The first report
Start with a question a report can answer. A fictional repair shop wants to compare completed jobs by service category for one month. Its table has one row per job, with a job ID, completion date, category and billed amount. Keep that grain visible as you work: counting rows should count jobs, while summing the amount should count revenue.
In Data Studio, connect a data source, create a report and add a table. Put category in the dimension and billed amount in the metric, using Sum. Add a count of job IDs as a separate metric. A dimension groups rows; a metric aggregates the rows in each group. Summing an ID can produce a number, but it does not measure how many jobs the shop completed.
Our Studio basics practice covers this distinction in a bounded browser simulation. It is part of the Looker and LookML course, rather than a separate full Data Studio course. It does not connect to a Google account or publish a real report.
Dates and calculated fields
Use a complete date field as the report’s date-range dimension. Then add a date control and confirm which charts respond to it. A chart with its own fixed date range can disagree with a neighboring chart even when both use the same source. Practice the date boundary in Studio date ranges.
For an average value per job, divide total billed amount by the number of distinct job IDs at the same aggregation level. Before using the result, check a category with repeated IDs and one with no completed jobs. A row-level calculation and a ratio of aggregated metrics can answer different questions. The official calculated-field reference explains which formulas belong at the source and which belong in an individual chart.
Blends and duplicate totals
A blend joins data for a report. Write down what one row represents in each input, then choose the join key. If a category summary has one row per category but its target table has several rows per category, joining only on category can multiply amounts. A matching field name does not establish a one-to-one relationship.
Check the total before and after the blend. Add the date to the key when the intended grain is category and month, or aggregate the repeated side before joining. Keep the original source total beside the blended one until they reconcile.
Sharing and credentials
Sharing a report and authorizing its underlying data are separate choices. Owner’s credentials allow the source to query under the credential owner’s authorization. Viewer’s credentials require each viewer to have the necessary data access. The available choices depend on the connector and organization policy.
In Studio credentials practice, choose the appropriate boundary for a fictional audience. In a real report, test it as an intended viewer before distributing the link. The simulator models the choice; it does not create Google permissions.
Further practice
For reusable warehouse metrics, views and Explores, take the Looker and LookML course. If the names are confusing, read Looker vs Looker Studio. A Looker dashboard guide explains the model-backed workflow separately.