Persona Graph System
An IBM UX research platform for managing personas, mapping task journeys, and running research campaigns β all connected in a Neo4j graph.
π€ Personas
Create and edit persona profiles with skills, tools, pain points, fears, and relationship networks.
π Relationships
Create typed relationships with directional strength scoring between personas, teams, skills, and tools.
π· Nodes
Manage Skills, Tools, Scopes, and Content Sources with category filtering.
πΈοΈ Graph
Interactive D3 graph with sibling highlight, relationship weight circles, and relationship flyout editing.
πΊοΈ Task Journeys
Collaborative swimlane editor β map sub-tasks across actors with fears, confidence, pains, and content sources per actor.
π Query
Custom Cypher queries with history, result visualisation, and graph exploration.
Current Features
- Rich Persona Profiles: Skills, tools, scopes, pain points, fears, content sources β all as first-class graph nodes
- Team Management: Teams with shared skills, tools, and scopes; team-to-persona relationships
- Directional Relationship Strength: Per-end scoring (-3 to +3) with context-sensitive labels (expertise/importance, proficiency/reliance, etc.)
- Task Journey Editor: Real-time collaborative swimlane with per-actor attitudes, content sources, and graph promotion
- Research Campaigns: Invite participants to complete journey boards; aggregate findings; promote pains, fears, and confidence to graph
- Graph Visualisation: Sibling highlight for shared content sources, relationship weight tip circles, tri-state legend filtering
- Category-based Node Management: Skills and tools organised into 12-category taxonomies with filterable dropdowns
- Inline Content Source Creation: Researchers can add unlisted sources directly from the journey map
- Dark / Light Theme: Carbon Design System with full dark mode support
- RESTful API: Full CRUD with interactive Swagger docs at /docs
Planned Enhancements
- Persona β Airtable Campaign Bridge: PII-safe audience selection β researchers pick criteria, backend resolves matching participants without exposing contact data
- PII Strategy: GDPR-compliant, IBM-sanctioned approach to research subject contact data β currently in proposal stage for CIO review
- Relationship Attribute Visual Encoding: Edge appearance driven by from/to weight values once sufficient data is populated
- Data Normalisation: Schema cleanup pass β canonical model audit, duplicate consolidation, safe migration scripts
- Open Source: Plugin architecture to allow external contributors to add new research types β no decision made yet
- AI Portrait Generation: Royalty-free persona portraits via Stable Diffusion β deferred pending model quality improvements
- Natural Language Queries: LLM-powered conversational interface over the graph
- Persona Similarity Analysis: ML clustering to identify similar personas across teams and organisations
Interaction
| Ctrl+Shift+V | Toggle dark / light theme (all pages) |
| Graph page |
| Click relationship | Open relationship properties & strength editor |
| Click shared node | Blue ring highlights all sibling copies |
| Double-click node | Make node the graph focus |
| Tab | Autocomplete node search suggestion |
| Enter | Select suggested node |
| Esc | Close dropdown / clear sibling rings |
| Journey / Tasks |
| Ctrl/β+click | Multi-select sticky cards |
| Esc | Close sticky detail panel |