Structured Narratives for Cultural Heritage Storytelling

Structured Narratives for Cultural Heritage Storytelling is a dataset of structured narratives built from the collection catalogue of the Museo Egizio in Torino. Each entry connects a cultural heritage artifact with a scene-based story generated for storytelling experiences in cultural heritage contexts. 

The dataset contains 1200 stories, organized into four subsets: Full Pipeline Structured Narratives, Direct Prompting Structured Narratives, Story Potential Structured Narratives, and No Scaffold Structured Narratives. Each subsets includes 300 stories generated using three AI models, with each model producing 100 stories per sub-dataset. 

Each entry consists of a structured scene-based story containing both general story-level information and details about the individual scenes. Except for the Direct Prompting dataset, each story is accompanied by the narrative unit from which it was generated, namely the narrative event extracted from the catalogue fields. Each entry also preserves source catalogue information related to the specific artifact. In the Full Pipeline subset, entries additionally include a Polti-based classification of the narrative unit.

The dataset is intended to support the creation of narrative experiences in cultural heritage settings by providing scene-based structured stories suitable for adaptation into immersive and interactive formats across different media. Grounded in real artifacts and catalogue information from a major museum collection, the stories are relevant for cultural contexts aimed at creating experiences around Ancient Egyptian heritage. In the Full Pipeline Structured Narratives subset, Polti’s dramatic situations provide an additional dramatic framework that may support the design of more engaging narrative experiences.

The dataset was developed in the context of a system designed to support curators in identifying collection artifacts with strong narrative potential and creating structured stories around them. It addresses two practical challenges in cultural heritage storytelling: identifying, within large collections, the artifacts that are most suitable as central elements of narrative experiences, and reducing the curatorial effort and expertise otherwise required to transform catalogue information into coherent stories.

The dataset creation is a joint effort between Università di Torino and Politecnico di Torino and stems from a collaboration with the Museo Egizio in Torino in the scope of project “Cultural Heritage Active Innovation for Next-Gen Sustainable Society”, CHANGES.