How it fits together

Four building blocks make up everything in Odoo Dashboards. Once you know what each one is for, the rest of the app is easy to find your way around.

1. Odoo modelYour records, as Odoo stores them.sale.order
2. DatasetThe fields you picked from that model, ready to chart.Sale Orders
3. WidgetOne tile, chart or table that summarises a dataset.Total Sales
4. DashboardWidgets on a grid, with filters, shared with people.Sales Overview

Datasets: the data you work with

A dataset is a table of data built from one Odoo model. You choose the model (for example Sales Order) and the fields you need (order date, customer, salesperson, total). Fields of related records count too, such as the customer’s country or the salesperson’s team.

Datasets are made once and reused. Ten widgets on five dashboards can all read from the same Sale Orders dataset. If the dataset gains a column, every widget built on it can use that column.

Datasets are created in the Dataset Builder. Advanced users can also write a dataset in SQL, combine two datasets, or store a dataset’s results so large data loads faster. See Datasets.

Tip

A dataset can carry a permanent filter, such as “only confirmed orders”. Every widget built on it then sees only those rows, so you don’t repeat the condition on each widget.

Widgets: one question each

A widget (the builder calls it an analytics item) answers one question about a dataset: what were total sales?, how did sales move month by month?, who are the top customers? Each widget has:

• a visualization type: Tile, Sticky Note, Vertical Bar, Horizontal Bar, Line Chart, Area Chart, Doughnut, Funnel, Heatmap, Map, Table, Pivot Table or Matrix Report;
• a data source, which is one dataset or several joined together;
• a measure and dimension, such as sum of Total by Order Date, per month;
• its own filters, some fixed and some that viewers can change.

Widgets are built in the Dashboard Builder, one at a time, with a live preview beside the settings. Choosing a widget compares every type.

Dashboards: widgets on a page

A dashboard is a page of widgets on a 12-column grid. Every dashboard gets its own entry in the Odoo Dashboards menu, and its own access list: only the users and groups you add can open it.

Each viewer can drag and resize the widgets on their own screen. Their layout is saved for them only. A builder can publish one layout as the default everyone starts from.

Filters: three levels

Filters are where most of the power is, and where most questions come from. There are three kinds:

Filter	Set up in	Affects
Widget filters	The widget’s Filters tab	That one widget. They can be fixed (always applied, invisible) or shown as chips the viewer can change.
Global filters	The builder’s Global Filters	Several widgets at once, from a filter bar at the top of the dashboard.
Date Filter	Dashboard Settings → Enable Global Date Range Filter	The main date filter of each widget, from one date picker in the dashboard header.

The rule to remember: filters live on the widget, not on the dataset. Filters covers all three in detail. A global filter doesn’t add a condition to widgets by itself. It drives filter lines that already exist on each widget, and you pick which ones. A widget with no matching filter line ignores the global filter and shows a Not filtered badge, so viewers know.

How a number reaches the screen

When someone opens a dashboard:

1. The dashboard loads its layout and shows a placeholder for every widget.
2. For each widget, Odoo describes the query (which dataset, which columns, how to summarise them, which filters) and asks the Odoo Dashboards service for the SQL.
3. The SQL runs on your own database, against the dataset, and the result fills in that widget. Widgets load independently, so one slow table doesn’t hold up the tiles.

Results are kept in the browser for five minutes, so moving between dashboards stays fast. The Refresh button in the dashboard header fetches everything again.

Who sees which records

Dashboards read datasets directly from the database, so Odoo’s record rules and multi-company restrictions don’t apply to them. Anyone who can open a dashboard sees every row its datasets include. To limit what a dashboard shows, filter the dataset (for example by company), and share the dashboard only with the people who should see that data.