Concept map

Updated
View as text
Open in Confluence
Open in ChatGPT Open in Claude

A concept map differs from a mind map in one important way: the links are labelled, and any concept can link to any other. That turns a hierarchy into a network, and the labels carry most of the meaning.

It is the better choice when the relationships between ideas are the thing you are trying to understand.

#When to reach for one

  • Learning or teaching a domain, where the relationships matter.

  • Domain modelling before committing to entities.

  • Untangling a subject where everything seems to relate to everything.

#How to draw it in Capable

Route

Use when

Graphviz

Arbitrary graphs with labelled edges and automatic layout.

Mermaid

Smaller maps where the flowchart syntax is enough.

#A worked example

Concepts and the relationships between them, labelled:

Concept map

The source, which you can paste into a new diagram and edit:

digraph concepts {
  rankdir=TB;
  node [shape=ellipse, fontname="Helvetica"];
  edge [fontname="Helvetica", fontsize=10];

  page -> macro [label="contains"];
  macro -> diagram [label="renders"];
  diagram -> source [label="is stored as"];
  source -> custom_content [label="lives in"];
  custom_content -> page [label="attached to"];
  diagram -> language [label="is written in"];
  language -> renderer [label="chooses"];
  renderer -> image [label="produces"];
}

#A few things that catch people out

  • Label every edge. An unlabelled concept map is a mind map with extra lines.

  • Read each link aloud as a sentence: subject, verb, object. If it does not read, the link is wrong.

  • Cross links between branches are the valuable part; a map with none is a hierarchy.

Generate first. Organise second.