Data Validation. Visuals, UX, ease of use.
Filtering, pagination and search features work as expected.
Calculations for graphs are the same on the dashboard as on the individual feature page.
Different viewports. Does it break if the screen gets smaller, load time.
Tested quite a few in my first testing role. Most of the problems I found were when applying numerous filters on a graph, the graph would often break out of its area or the data in the graph would start to overlap particularly on bar graphs. Also validate the data against backend.
What else could you add to the conversation? How do you test a dashboard?
Itโs similar to a web app in that itโs a combination of front and back ends, so Iโd want to know what the scope is for the testing. Assuming itโs everything, on top of all the good things people have already said:
Whatโs the delay between the world changing, e.g. someone buying something from our online store, and it showing up in the dashboard? What should the delay be?
How does it cope with months having different lengths?
Is it possible for data to be late, and if so how does it cope with that / show it? For instance, in a previous job we were billing phone calls after they happened. What happens when weโre told in April about a phone call that happened in March? Is it indistinguishable from an April phone call? Is it included at all?
But the main thing I would want to know is: what decisions or actions is this dashboard meant to support? (Note that this drags in a definition of expected users, who will be making the decisions or performing the actions.) Then I would test how well it did that. All the rest is details - the technology, the graphic design etc. (A bit of relevant blogging: Visualisations and the stories behind them โ Random Tech Thoughts)