An ontology for representing and analyzing analytical provenance information from visualization tools.
DIVA provides and ontology and a thesaurus to:
- Model user interactions with visualization tools
- Capture exploration paths and patterns
- Enable semantic analysis of visual analytics process
- Support reuse and comparison of analysis strategies
The DIVA namespace is http://ns.inria.fr/diva/
See the DIVA ontology repository.
To demonstrate the usage of DIVA ontology, we lifted system logs describing analytical provenance information from the eSTIMe visualization system, a multi-screen visualization platform designed for exploring urban mobility data.
The system logs describe the exploration processes of six users, varying from 31 to 207 steps (i.e., performed actions within the visualization). Each log was recorded as a JSON array, structured sequentially to represent user interactions and the corresponding visualization states. Each log entry includes metadata such as timestamp, action type, target, view context, and descriptors of the resulting artifacts (e.g., visualizations or datasets). We used the rdflib library in Python to transform this structured JSON data into an RDF KG.
A set of competency questions were implemented through SPARQL queries and interactive visualization tools, supporting visual exploration of analytical provenance information.
See the Case Study repository.
diva/
├── ontology/ # DIVA ontology files
│ ├── diva.ttl # Core OWL ontology in Turtle format
│ └── thesaurus.ttl # Controlled vocabulary
│
└── case-study/ # Example implementation
├── data/ # Input and output data used in the case study
├── lifting/ # Data conversion scripts
├── sparql-examples/ # Example SPARQL queries and competency questions
└── vis/ # Visualization components
See the LICENSE file.