Narrative Classification
The Narrative Classification into Tags functionality in Cardus.AI allows you to organize narratives based on classes explicitly defined by you. Unlike categories applied to interviews, tags are assigned directly to narratives, enabling cross-sectional and comparable analysis of experiences, situations, and patterns reported throughout the project.
Key Features
- User-Defined Tags: Freely create tags that represent the types of situations, tensions, or practices you want to observe in narratives
- Classification Based on Real Examples: The classifier uses real narratives (chosen by you) from the project as references, avoiding generic or abstract labels, and helping with the interpretation of the specific context
- Direct Application to Narratives: Tags are assigned at the narrative level, not the interview level
- Multiple Tags per Narrative: A single narrative can belong to more than one tag, reflecting the natural overlap of organizational experiences
- Batch Processing: Classify large volumes of narratives simultaneously in a background process
- Classification History: Each execution is recorded, allowing auditing, comparison between versions, and future reprocessing
- Dashboard Integration: Tags feed visualizations and comparative analyses throughout the project
- Tag Findings: For each tag, Cardus.AI automatically generates a set of synthetic findings that summarize recurring patterns observed across the classified narratives. These findings do not replace reading the narratives, but help guide analysis and comparative exploration.
Benefits
- Organize Large Volumes of Narratives: Transform an extensive qualitative corpus into comparable and navigable sets
- Compare Patterns Between Groups: Analyze the distribution of tags by department, position, period, or other segments defined in the interviews
- Increase Analytical Consistency: Reduce interpretative variations by applying classification criteria systematically, perfect for comparing before and after an organizational intervention
- Support Synthesis Without Excessive Simplification: Structure the analysis without reducing narratives to single metrics or normative judgments
How to Use
- Define Tags: Create the classes you want to use to classify narratives
- Select Examples: With the help of a semantic search tool, choose representative narratives for each tag to guide the classifier
- Run Classification: AI applies tags and establishes a confidence value between 0 and 1 for each narrative
- Review and Re-run: Adjust the minimum confidence value acceptable to you, so you decide how rigorous the classification will be
- Explore Results: Use tags to filter narratives, compare distributions, and deepen analysis in the Dashboard
Note: Ask for help in the Narratives page chat, as it can assist you in defining classes and answering questions.