Data Studio reports and dashboards

Friday, October 30, is the first visual session. Lena wants a dashboard for the forty store managers, ready for a planned launch on December 1, that answers the questions they ask her every week: how are my sales this week, how does that compare with last year, are returns rising, and which categories are driving the change? Ravi and Mateo build the first version in Data Studio, Google's free reporting tool, and plan the governed version in Looker, covered in lesson 9.

Data Studio, Looker Studio and Looker

Names first, because they confuse everyone. Google's free reporting tool was launched as Data Studio, renamed Looker Studio in 2022, and renamed back to Data Studio on April 10, 2026. The paid tier, formerly Looker Studio Pro, is now Data Studio Pro. Existing reports, data sources and users moved over automatically. Many exam questions, practice tests and articles still say Looker Studio; read them as meaning Data Studio. Looker, without "Studio", is a different product: Google's enterprise business intelligence platform built on a semantic model written in LookML.

The Associate Data Practitioner exam asks you to compare Looker and Looker Studio, now Data Studio, and the comparison comes down to governance. In Data Studio, each report or data source can define metrics its own way, which is fast and flexible but lets definitions drift. In Looker, metrics are defined once in the model and every dashboard inherits them. A practical rule: use Data Studio when speed and flexibility matter and the audience is small or the analysis is exploratory; use Looker when many people across the organization must see the same number defined the same way.

Data sources and connectors

A Data Studio report reads data through a data source, which uses a connector to reach a system such as BigQuery, Google Sheets, Google Ads, Cloud SQL or a file upload. For BigQuery, the data source can point at a table, a view or a custom query. Ravi points it at the shared view curated.v_store_daily from lesson 5, so the dashboard uses the same definition of net sales as every other report.

Two data source settings matter a great deal. The first is data credentials: with owner's credentials, the report queries BigQuery as the person who created the data source, so viewers do not need their own access to the data; with viewer's credentials, each viewer's own BigQuery permissions apply, so they see only what they are allowed to see and their queries are billed to their own project setting. The second is data freshness: Data Studio caches query results for a set period, by default 12 hours for BigQuery, and viewers see cached data until it expires or someone refreshes. A daily dashboard loaded at 06:00 can use a long freshness; an operational dashboard may need 15 minutes.

Owner's credentials are convenient, but they turn the data source owner into a gateway to the data: anyone who can view the report sees whatever the source returns. Lesson 10 shows how Birchline combines owner's credentials with a view that limits what is exposed, and when viewer's credentials are the safer choice.

Fields, calculated fields and the ratio trap

Each column in the data source becomes a dimension, something to group or filter by such as region or sale_date, or a metric, something to aggregate such as net_sales. A calculated field adds a new dimension or metric with a formula, either in the data source, where every report using it shares the definition, or in a single chart.

Mateo creates a return rate field and walks straight into the most common calculated-field mistake: averaging ratios. If each row already holds a daily return rate, averaging those rates gives every day equal weight, whatever its sales. The correct return rate for a period is the total returns divided by the total sales, with both sums taken first. In Data Studio that is SUM(refunds) / SUM(net_sales), not AVG(daily_return_rate).

# Daily figures for one store: (net sales, refunds), in dollars
days = [(12000, 480), (11000, 495), (3000, 450), (14000, 560)]

avg_of_ratios = sum(r / s for s, r in days) / len(days)
ratio_of_sums = sum(r for s, r in days) / sum(s for s, r in days)

print(f"Average of daily return rates: {avg_of_ratios:.4f}")
print(f"Total refunds / total sales:   {ratio_of_sums:.4f}")
print(f"Per 100 dollars of sales: {avg_of_ratios * 100:.2f} vs {ratio_of_sums * 100:.2f}")

Average of daily return rates: 0.0688
Total refunds / total sales:   0.0496
Per 100 dollars of sales: 6.88 vs 4.96

The quiet third day, with sales of 3,000 dollars and refunds of 450, has a rate of 0.15, and averaging lets it pull the result up to about 0.0688. The true rate for the four days is 1,985 dollars of refunds on 40,000 dollars of sales, or about 0.0496. The dashboard uses the ratio of sums, defined once in the data source.

Designing the dashboard

A dashboard is not a collection of every chart that could be drawn. Ravi starts with the managers' four questions and designs one page that answers them in order, from the top left, where eyes go first, to the bottom right.

The design follows a few principles worth adopting. Scorecards at the top answer "how are we doing?" with one number each and a comparison, such as the same week last year. Charts below give context: a line chart for change over time, a sorted bar chart for comparing categories. A table at the bottom gives detail for anyone who wants to look up a specific product. Controls let managers change the date range and choose their store; a date range control applies to the data source's date dimension. A line at the top states when the data was last updated, so nobody mistakes yesterday's figures for today's.

Data Studio offers many chart types and styling options, and restraint helps. The team's guide limits each page to one accent color for emphasis, avoids 3D and pie charts with more than a few slices, starts bar chart axes at zero, and labels units in titles, such as "Net sales (dollars)". Lesson 11 covers chart choice and honest presentation in more detail.

Blending, filters and interactivity

Blending combines data from more than one data source in a chart, joined on shared dimensions, for example store sales from BigQuery with store targets kept in a Google Sheet. Blends support several join types and are handy for quick combinations. For anything used widely, Ravi prefers to do the join in BigQuery, in a view, so the logic is visible, testable and reusable outside the report.

Filters can be set on a chart, a page or the whole report. Cross-filtering lets a viewer click a bar, such as the footwear category, and filter every other chart on the page to footwear. Parameters let viewers enter a value, such as a target growth rate, that calculated fields use. Drill-down lets a chart move from region to store to category. These features make one dashboard serve many questions without many copies.

Performance and cost

Each chart in a Data Studio report sends its own query to BigQuery when the cache is empty. A page with twelve charts opened by forty managers each morning can mean hundreds of queries. Four habits keep that fast and cheap: point the report at small, pre-aggregated data such as a materialized view or a summary table rather than raw events; filter on the partition column through the date range control; set a sensible data freshness so the cache absorbs repeated views; and consider BI Engine, which keeps frequently used data in memory and speeds up Data Studio queries considerably.

Data Studio Pro adds features that matter for an official dashboard: content owned by the organization rather than an individual's account, team workspaces, Google Cloud project-level permissions and audit logs, AI features through Gemini, and support. Because the store manager dashboard must outlive any one employee, Ravi plans to build it in a Pro team workspace.

Sharing the first version

Ravi shares the draft with Lena and two pilot store managers as viewers, with link sharing off. Feedback arrives within an hour: one manager wants the comparison to last year shown as a change per 100 dollars of last year's sales, the other wants to see weekdays and weekends separately because her store is in a tourist area. Both changes take minutes, which is exactly what Data Studio is good at. Governed, organization-wide definitions are the subject of the next lesson.

Before the pilot, Mateo tests the dashboard on a phone, because many store managers will open it on the shop floor. Scorecards stack well on a small screen, but the bar chart of categories is too cramped, so he shortens the category labels and limits the chart to the top eight categories with an "All other" bar.