license: mit
task_categories:
- summarization
- text-generation
language:
- en
tags:
- writing
- storytelling
- narrative-analysis
- plot-structure
pretty_name: PlotStructureSummarization-Generation
size_categories:
- n<1K
Dataset Card for PlotStructureSummarization-Generation
This dataset consists of manually curated plot summaries across various media types, such as TV episodes, movies, books, and historical events. Each instance follows a structured format outlining the title, synopsis, plot structure, and key character archetypes. It is designed to enhance models' abilities in summarization and text generation, focusing on narrative frameworks, including classical storytelling structures like the Hero’s Journey, Quest, and others. This dataset serves both as a tool for research in narrative theory and as a training resource for models focused on structured narrative generation.
Dataset Details
Dataset Description
The PlotStructureSummarization-Generation dataset provides detailed, structured plot summaries focusing on narrative structures and character archetypes. It is designed to help models understand and generate structured plot summaries by identifying common storytelling techniques and structures. It is particularly useful for tasks in summarization, text generation, and narrative analysis, aimed at better understanding how stories unfold across various types of media.
- Curated by: Manually Curated by Robert McNarland, McNarland Software Consultants
- Funded by [optional]: None
- Shared by [optional]: None
- Language(s) (NLP): English
- License: MIT
Dataset Sources [optional]
The dataset pulls from a wide variety of publicly available narratives, including TV episodes, movies, books, and historical events. The summarization process focuses on highlighting classical plot structures and universally recognizable character archetypes.
- Repository: [More Information Needed]
- Paper [optional]: [More Information Needed]
- Demo [optional]: [More Information Needed]
Uses
Direct Use
This dataset is suitable for several types of tasks:
- Text Summarization: Training models to summarize narrative plots concisely while retaining key structural elements.
- Text Generation: Developing systems that generate structured plot summaries or create new stories based on classical narrative frameworks.
- Narrative Understanding: Research and development of AI systems focused on recognizing and reproducing classical story structures, such as the Hero’s Journey or Quest.
- Creative Writing Assistants: Assisting in the creation of structured stories by leveraging predefined plot archetypes.
Out-of-Scope Use
This dataset is not intended for tasks outside narrative generation and summarization. It should not be used for unrelated tasks such as data classification or any numeric-based data processing. The dataset is also not designed to generate culturally insensitive content or plots that might propagate biases without oversight.
Dataset Structure
Each instance consists of the following fields:
- title: The title of the media or event.
- author: Type of the media (e.g., Movie, TV Episode, Book, Historical Event).
- synopsis: A structured summary detailing the exposition, rising action, climax, and resolution.
- plot_structure: The narrative structure used (e.g., Hero's Journey, Quest, Irony).
Dataset Creation
Curation Rationale
This dataset was created to support AI research in storytelling and narrative generation. By summarizing plots into a structured format, it allows models to better understand and generate coherent narratives. This structured approach can lead to more accurate text generation, especially for tasks involving storytelling, plot summarization, and creative writing.
Source Data
Data Collection and Processing
Data was collected manually from publicly available media, including books, TV shows, movies, and historical narratives. Each entry was annotated with its plot structure and relevant character archetypes to ensure consistency across instances.
Who are the source data producers?
The data consists of well-known stories and publicly available media. Summaries are derived from common knowledge and publicly accessible sources.
Annotations [optional]
Annotation process
Each plot summary was manually curated to fit a predefined structure. The annotation focused on narrative components, ensuring that the plot and characters aligned with classic storytelling frameworks like the Hero’s Journey or Freytag's Pyramid. Each summary was reviewed for accuracy, coherence, and adherence to the structure.
Who are the annotators?
The annotators are experienced in summarization and have a background in narrative analysis. They manually curated and structured each entry to fit a predefined format.
Personal and Sensitive Information
This dataset does not contain any personal or sensitive information. It focuses entirely on publicly available media plots.
Bias, Risks, and Limitations
Recommendations
Bias
The dataset may reflect biases inherent in the selection of media types or in the interpretation of plot structures. While efforts were made to include a variety of media, the selection is not exhaustive, and some narrative styles (especially non-traditional or experimental forms) may not be well-represented.
Risks
Since the dataset is manually curated, there may be some subjective interpretations of plot importance and structure. Additionally, users should be aware that the dataset focuses on conventional storytelling forms and might not capture more nuanced or culturally specific narrative styles.
Limitations
This dataset focuses on classical plot structures and may not be applicable to non-traditional narratives or highly experimental storytelling methods. It may also not be suitable for tasks requiring detailed or granular analysis of subplots, themes, or character development outside of traditional structures.
Citation [optional]
BibTeX:
[More Information Needed]
APA:
[More Information Needed]
Glossary [optional]
[More Information Needed]
More Information [optional]
Further resources or tools that might be useful for users of the dataset. This might include links to relevant books, articles, or tools that aid in narrative analysis.
Dataset Card Authors [optional]
Authors: Robert McNarland, McNarland Software Consultants
Dataset Card Contact
[More Information Needed]