A tool for analyzing book structure using NLP techniques. See the forest instead of the trees.
Story Inspector consists of the following components:
- An ETL Tool (Extraction, Transform, Load) that takes books in various formats (HTML, ePUB, PDF, etc.) as input and produces a StoryDOM (Document Object Model) file breaking down the book in its structural elements (parts, chapters, and paragraphs).
- An Annotation Engine that uses NLP to produce metadata from StoryDOM files, enriching the files with metadata annotations. Examples of metadata include:
- Word Count
- Characters
- Locations
- Sentiment Score
- Emotion Score
- Etc.
- A Visualization Tool for authors to interpret the results.
- A Web Visualization Tool is hosted at https://github.com/taciano-perez/story-inspector-web.
Limitations
- Story Inspector cannot analyze sentiment nor extract named entities (locations, characters) from sentences larger than 250 words.
- Story Inspector cannot yet rename or delete characters or locations via the UI