6.4 Building, Modifying & Sharing Looker Dashboards

6.4 Building, Modifying & Sharing Looker Dashboards

Core Focus: Exam section 2.2 is titled "Visualize data and create dashboards in Looker." Section 6.3 explained the LookML model behind Looker. This section covers what an analyst does on top of it: building dashboards from Explores, changing them, sharing and scheduling them safely, and making small LookML edits that change what users see.

---

Two Kinds of Looker Dashboards

You can start with a user-defined dashboard and use Get LookML to copy its definition into a project when it needs version control.

---

Building a Dashboard from an Explore

1. Open an Explore (for example, Orders & Fulfillment Analysis).
2. Pick fields: choose dimensions (such as Order Created Month, Region) and measures (such as Total Gross Revenue).
3. Filter: add filters such as Order Status is Complete and a date range.
4. Run and visualize: choose a column chart, line chart, table, single value, or map.
5. Save: save the query to a dashboard as a new tile, or save it as a Look (a saved, reusable report) and add the Look to dashboards.

Because every tile is generated from the LookML model, two tiles that use Total Gross Revenue always use the same definition. That consistency is the main reason to use Looker rather than per-report calculated fields.

Tiles, Filters, and Interactivity

• Query tiles hold their own query; Look-linked tiles show a saved Look and change when the Look changes.
• Text and Markdown tiles add headings and explanations.
• Dashboard filters apply to every tile you map them to (choose the field each tile should filter on), so one date filter can drive the whole dashboard.
• Cross-filtering lets a viewer click a bar or row to filter the other tiles by that value.
• Drilling opens the detail rows behind a number, using the drill_fields defined in LookML.
• Settings control auto-refresh, the dashboard time zone, and whether tiles run when the dashboard loads.

---

Modifying Dashboards

In Edit mode you can move and resize tiles, change a tile's visualization or query, add and remove filters, and rename the dashboard. Changes to a user-defined dashboard take effect when you save. Changes to a LookML dashboard follow the same development workflow as other LookML.

---

Sharing, Scheduling, and Access

Looker separates content access (who can see a dashboard in a folder) from data access (which models, Explores, fields, and rows a user can query).

Because queries run under the viewer's Looker permissions and access filters, one shared dashboard can safely show each regional manager only their region.

---

Making Simple LookML Changes

Exam section 2.2 also asks you to "manipulate simple LookML parameters to modify a data model." In LookML, a parameter is any key: value setting on a field, view, Explore, or model. The ones you are expected to change are simple:

measure: total_gross_revenue {
  label: "Gross Revenue (USD)"
  description: "Sum of sale price for completed order items"
  type: sum
  sql: ${sale_price} ;;
  filters: [status: "Complete"]
  value_format_name: usd
  drill_fields: [id, created_date, sale_price]
}

dimension: internal_cost_code {
  hidden: yes
  sql: ${TABLE}.cost_code ;;
}

LookML also has a field type called parameter, which adds a user-selected input (for example, a "metric selector") that other fields read with Liquid templating. Recognize it, but most exam items focus on the field settings above.

The Development Workflow

1. Turn on Development Mode. You now work on your own Git branch; other users still see production.
2. Edit the view, model, or dashboard file.
3. Click Validate LookML to catch errors.
4. Commit your changes (and open a pull request if your team requires review).
5. Deploy to Production so everyone sees the change.

---

Common Exam Traps

• "I changed the LookML, but nobody sees it." Changes in Development Mode are visible only to the developer until they are committed and deployed to production.
• Sharing data versus sharing content: Moving a dashboard into a shared folder does not grant access to a model the viewer cannot query; data access comes from roles and access filters.
• Copying formulas into tiles: Repeating a revenue formula in table calculations across dashboards recreates metric chaos. Define it once as a LookML measure.
• Looker vs. Looker Studio: Governed dashboards on a semantic model point to Looker; quick, free reports from many sources point to Looker Studio (Section 6.2).