Workflow

The challenge in our sector — cities and buildings — is fragmented, multi-disciplinary workflows. This means they're hard to see, so hard to understand, and so hard to improve on.

Visualization

Giraffe believes in visualisation. We love John Snow's Cholera map. We love Edward Tufte.

Giraffe is — in one way — just a visual version of a spreadsheet on a map.

Structured text

In addition to visualization, Giraffe loves text, especially structured text. Particularly JSON.

The Giraffe native format is GeoJSON (geographic JSON). It is human readable. It is machine readable. It doesn't make the mistakes of old formats that tried to save bytes by truncating headings or compiling down to binaries. It is not trying to create a commercial advantage by being obscure.

It is a baggy, beautiful, capacious, elegant format. (LLMs love JSON too — because LLMs love text.)

Giraffe is on the computer because the computer can turn the text into a visualization. The data and its visualization feel like one thing — because the computer converts one to the other.

Coupling data to visualization

So naturally, when we built an app to solve the hard problem of thinking about workflows, we built a data structure tightly coupled to its visualization. Nothing amazingly novel, but with some nice connections to Giraffe.

What comes next

The next steps seem clear. The JSON structure explaining the workflow will be fed in a structured way to an LLM that will execute it.