Translate
SKOS Taxonomies
Rich taxonomies are the key to successfully harnessing Tridion’s Semantic Tagging and Search capabilities. But creating taxonomies can be time consuming and costly.
People use ChatGPT to write text. We can also use ChatGPT to build and visualise taxonomies. ChatGPT already knows SKOS (Simple Knowledge Organization System). For example:
Generate a taxonomy in SKOS format describing the relationship between concepts relating to Computer Hardware |
ChatGPT responds with:
|
|
This can then be imported into Taxonomy Manger as the basis of a new taxonomy.
Visualisation
What if you want to visualise this taxonomy quickly? The trick is to teach it how to format the text. Let’s take markmap (https://markmap.js.org) – a tool used to visualise mindmaps from markdown.
We can teach ChatGPT how markmap works, and then ask it to build a taxonomy on a subject in that format:
|
|
ChatGPT will come back with an HTML formatted list:
|
We can ask ChatGPT to format it as code:
Escape the mind map as a code block |
Chat GPT will give us:
# Car Manufacturing## Parts### Engine### Chassis### Transmission### Brakes## Assembly### Body### Painting### Testing## Logistics### Shipping### Inventory Management## Suppliers### Raw Materials### Component Manufacturers |
This mind map isn’t very extensive, so we can ask ChatGPT for more and it will provide a more comprehensive result:
Generate a longer mindmap on Car Manufacturing |
This results in the following:
|
|
We can also ask ChatGPT to translate this into other languages:
Translate the concepts to Dutch |
Resulting in:
|
|
Now, that ChatGPT knows what to do, you can generate a taxonomy for a different domain like this:
Using markdown mind map diagram language markmap make a mind map to describe the relationship between concepts used in Healthcare. Generate a larger number of concepts. |
Resulting in:
|
|
The output can be visualised using https://markmap.js.org/repl
We can also make it look fancy using Mermaid:
|
|
Which results in:
mindmap root((Car Manufacturing)) Design Concept ::icon(fa fa-lightbulb) Engineering ::icon(fa fa-cog) Prototyping ::icon(fa fa-object-group) Production Assembly ::icon(fa fa-industry) Quality Control ::icon(fa fa-check-circle) Supply Chain ::icon(fa fa-truck) Marketing Branding ::icon(fa fa-bold) Advertising ::icon(fa fa-bullhorn) Sales ::icon(fa fa-dollar-sign) Technology Automation ::icon(fa fa-robot) Electrification ::icon(fa fa-bolt) Connectivity ::icon(fa fa-network-wired) Regulations Safety ::icon(fa fa-shield-alt) Emissions ::icon(fa fa-smog) Fuel Efficiency ::icon(fa fa-gas-pump) |
Paste this into https://t.co/inHUSQcenB, which will result in the following diagram:
ChatGPT can be a valuable tool in creating taxonomies for a particular industry, and for visualising them quickly in different manners. Clearly some caution needs to be applied to the results, which may not always be accurate. A Knowledge Manager should be able to audit the results and correct those that do not make sense.
Couple this with Tridion’s auto-tagging capabilities, this allows customers to quickly get started in classifying their content based on their industry.
Credit to https://twitter.com/Claybight for ChatGPT prompt inspiration.