Use an AI Assistant to create Custom Visualisations

nodegoat offers three built-in modes to visualise your data: spatial, social, and temporal. Due to the flexibility of these modes of visualisations you can generate complex visualisations that combine spatial, social, and temporal perspectives within the built-in options (e.g. diachronic developments of spatial networks).

To go beyond these visualisation modes you can make use of the 'Embed' feature that allows you to use a custom visualisation within your nodegoat environment. You can use this feature for custom visualisations programmed by yourself, or you can interact with an LLM in order to generate a custom visualisation. For both scenarios the documentation section provides useful pointers.

In this Guide we walk you through the steps of generating a custom nodegoat visualisation based on an interaction with an LLM.

In order to do this, go to Data and click the   icon on the right side of the visualisation buttons. This opens the 'Visualisation Settings' dialog. Go to the fourth tab 'Visual Settings'.   Click the tab 'Embed'. Click '[+Full Description]' to find the context that describes the technical aspects for a nodegoat visualisation and shows an example that uses all options. Click this description to copy this.

Go to any AI assistant and paste this text and add a prompt like 'Use this context to create a custom visualisation that presents the included Objects both in a Jaccard index similarity matrix as well as in a Jaccard index similarity network cluster. Both visualisations should be shown side by side in an interactive manner so that if a user hovers over one object it is also highlighted in the other visualisation. Include relevant options as adjustable filters.'

Copy the generated code and paste it in the 'Script' field. Click 'Apply Settings'. You now see a fourth option in the toolbar section containing the visualisation buttons. Click the   icon to run the custom visualisation.

This is just one example of the endless amount of possibilities that can be explored using this functionality.