[vomitcuddle]

A dashboard is a web app displaying aggregate data - usually on something like a digital signage display at a company's office - often includes 'vanity metrics' like web traffic data, Git commits, social media feeds, CI builds, etc.

[jldugger]

To expand upon that, it's aggregating a variety of data sources, generally via disparate APIs, rather than creating its own database backed API.

So while most framework models are oriented around their own ORM, a dashboard is probably better off modeling a set of translators, and providing tools for adding new translators. Translators can take a random API, and convert into a normalized data type the rest of the framework can understand. And a good community will have a lot of these translators available.

Additionally, the dashboard framework can provide a variety of standard visualizations for the translators to send normalized data. Time series, histogram, pie chart, text, etc. And provide an intelligent layout engine for arranging a collection of these on a screen. Dashing, for example, allows a user to re-arrange dashlets.

tl;dr -- it _is_ a webapp, but the more limited scope allows for better tooling and your valuable engineers can focus the more productive uses of their time, like writing new visualizations or, more likely, new translators.