University of Bologna logo
Universidad Politecnica de Madrid logo

Deliberation Knowledge Graph

Connecting Deliberative Process Datasets and Ontologies for Enhanced Analysis and Understanding

Deliberation Knowledge Graph

Connecting Deliberative Process Datasets and Ontologies for Enhanced Analysis and Understanding

About the Project

The Deliberation Knowledge Graph project aims to connect various deliberative process datasets and ontologies into a comprehensive knowledge graph. This project provides a unified framework for analyzing deliberative processes across different platforms and contexts, from formal parliamentary debates to citizen participation initiatives.

By integrating diverse datasets and ontologies, the Deliberation Knowledge Graph enables researchers, policymakers, and citizens to explore the connections between different deliberative processes, identify patterns, and gain insights into how deliberation works across various contexts.

Authors

SV

Simone Vagnoni

CIRSFID, University of Bologna - OEG, Universidad Politecnica de Madrid

University Profile

Datasets

The project integrates several datasets related to deliberative processes from various sources:

Featured Dataset: European Parliament Debates

The European Parliament Debates dataset contains structured representations of plenary session debates, converted from verbatim HTML reports into both JSON and RDF formats aligned with the Deliberation Ontology.

This dataset captures the rich deliberative processes of the European Parliament, including speeches, topics, participants, and their political affiliations.

Data Structure

{
  "@type": "deliberation:DeliberationProcess",
  "deliberation:identifier": "ep_debate_YYYYMMDD",
  "deliberation:name": "European Parliament Debate",
  "deliberation:hasTopic": [...],
  "deliberation:hasParticipant": [...],
  "deliberation:hasContribution": [...]
}
European Parliament in session

Decide Madrid

Citizen proposals and comments from Madrid's participatory democracy platform, enabling citizens to propose, debate, and vote on city initiatives.

DeliData

A dataset for deliberation in multi-party problem solving, containing structured conversations aimed at reaching consensus on complex issues.

EU Have Your Say

Feedback and initiatives from the European Commission's public consultation platform, where citizens and stakeholders can provide input on EU policies.

Habermas Machine

Data from a deliberative democracy experiment inspired by Jürgen Habermas's theories of communicative rationality and deliberative democracy.

Decidim Barcelona

Data from Barcelona's implementation of the Decidim participatory democracy platform, showcasing citizen engagement in urban planning and policy development.

US Supreme Court Arguments

Transcripts of oral arguments before the United States Supreme Court (2017-2021), providing insights into legal deliberation at the highest judicial level.

Deliberation Ontology

The core ontology of this project, which integrates concepts from the ontologies listed below:

Deliberation Ontology high-level diagram

Ontology Mappings

The mappings.owl file contains explicit mappings between the Deliberation Ontology and other standard ontologies like FOAF, Dublin Core, SIOC, AIF, and LKIF. These mappings enable interoperability and integration with existing semantic web resources.

Key mapping relationships include:

  • skos:exactMatch - Concepts that are exactly equivalent
  • skos:closeMatch - Concepts that are very similar but not identical
  • skos:relatedMatch - Concepts that are related but not equivalent
  • skos:broadMatch - The source concept is broader than the target
  • skos:narrowMatch - The source concept is narrower than the target

Relevant Ontologies

Ontology Available (rdf | xml | doc) Name Description Language Jurisdiction/Scope Reference / Links
DELIB rdf | doc DELIB Ontology Geared to capture essential elements of deliberative processes and participant interactions, emphasizing how contributions, interventions, and outcomes connect within a deliberation. en all OWL)
Documentation
AIF rdf | doc Argument Interchange Format A core ontology designed to represent and exchange argument structures, enabling interoperability between argumentation tools. Developed to capture premises, conclusions, and relations in structured debate. en all AIF (OWL)
Documentation
SIOC rdf | doc SIOC Ontology The Semantically-Interlinked Online Communities (SIOC) Ontology provides concepts for describing online community activities, including posts, forums, and user accounts. Extended with argumentation structures to support semantic linking of discussion threads and arguments. en all SIOC (RDF)
Documentation
FOAF rdf | doc FOAF Ontology Friend of a Friend (FOAF) describes persons, their activities, and their relations to other people and objects. Often used to model participant identities, organizations, and groups in social web contexts. en all FOAF (RDF)
Documentation
DC rdf | doc Dublin Core A set of vocabulary terms used to describe web resources (title, creator, subject, date, etc.). Widely used for metadata, providing consistent description properties (e.g., dc:title, dc:creator). en all DC (RDF)
Documentation
LKIF rdf | doc LKIF Ontology The Legal Knowledge Interchange Format (LKIF) covers concepts for norms, rules, and legal reasoning. Often used to integrate legal arguments with broader domain ontologies in e-government and e-participation. en eu LKIF (OWL)
Documentation
IBIS rdf | doc Issue-Based Information System A classic method for modeling and visualizing issues, ideas, and arguments in a structured format. Widely used for deliberative problem-solving and mapping various positions and solutions. en all IBIS (OWL)
Documentation
AKN xml | doc Akoma Ntoso A structured XML schema defining representations of parliamentary, legislative, and judiciary documents (also known as LegalDocML). Facilitates interoperability for legal and legislative texts across jurisdictions. en all Akoma Ntoso (OASIS)
Documentation
Metalex xml | doc Metalex An open XML interchange format for legal and legislative resources, enabling consistent structuring and referencing of sources of law, often used alongside Akoma Ntoso. en eu Metalex Homepage
CEN Standards
LegalRuleML xml | doc LegalRuleML A framework extending RuleML with legal rule modeling capabilities, enabling interoperable exchange of legal arguments, norms, and complex reasoning structures in a machine-readable format. en all LegalRuleML (OASIS)
Documentation

Data Conversion Tools

The Deliberation Knowledge Graph project provides conversion tools to transform various deliberation datasets into RDF format compatible with the deliberation ontology:

Dataset Conversion Script Input Format Output Format Description
EU Parliament Debates data/EU_parliament_debates/convert_json_to_rdf.py JSON RDF/XML Converts structured JSON representations of EU Parliament debates to RDF format
US Supreme Court Arguments data/US_supreme_court_arguments/convert_to_rdf.py CSV RDF/XML Converts US Supreme Court oral argument transcripts to RDF format
Decide Madrid data/decide_Madrid/convert_to_rdf.py CSV RDF/XML Converts Decide Madrid proposals and comments to RDF format
DeliData data/delidata/convert_to_rdf.py TSV RDF/XML Converts DeliData deliberation conversations to RDF format
Habermas Machine data/habermas_machine/convert_to_rdf.py Parquet RDF/XML Converts Habermas Machine deliberation experiment data to RDF format
EU Have Your Say data/EU_have_your_say/convert_to_rdf.py CSV RDF/XML Converts EU Have Your Say feedback and initiatives to RDF format
Decidim Barcelona data/decidim_barcelona/convert_to_rdf.py CSV RDF/XML Converts Decidim Barcelona proposals, meetings, and comments to RDF format

All conversion scripts follow a two-step process:

  1. Convert the source data to a JSON format aligned with the deliberation ontology
  2. Convert the JSON to RDF/XML format using the common convert_json_to_rdf.py script

This approach ensures consistency across all datasets and makes it easy to add new datasets in the future.

Ontology-Specific Graphs

Below are separate Mermaid.js graphs showing how each ontology's core classes and properties map to the Deliberation Ontology (Del). We have extracted only the relevant portion of the large integrated diagram, so you can view them in smaller chunks, per ontology.

1) DELIB Ontology Alignment

graph TD DEL[Deliberation Ontology]:::mainNode DELIB[DELIB Ontology]:::ontNode DP[DeliberationProcess] S[Stage] ACTIVITY[Activity] P[Participant] ORG[Organization] ROLE[Role] IR[InformationResource] A[Argument] OPINION[Opinion] VOTE[Vote] LS[LegalSource] LD[LegalDecomposition] LF[LegalFramework] LI[LegalInterpretation] DP -->|skos:closeMatch| DELIB_P[delib:Process] S -->|skos:closeMatch| DELIB_S[delib:Stage] ACTIVITY -->|skos:closeMatch| DELIB_A[delib:Activity] P -->|skos:closeMatch| DELIB_ST[delib:Stakeholder] ORG -->|skos:narrowMatch| DELIB_FP[delib:FormalParticipant] ROLE -->|skos:relatedMatch| DELIB_R[delib:Role] IR -->|skos:broadMatch| DELIB_I[delib:Information] A -->|skos:exactMatch| DELIB_AR[delib:Argument] OPINION -->|skos:exactMatch| DELIB_O[delib:Opinion] VOTE -->|skos:exactMatch| DELIB_V[delib:Vote] LS -->|skos:exactMatch| DELIB_LS[delib:LegalSource] LD -->|skos:exactMatch| DELIB_LD[delib:LegalDecomposition] LF -->|skos:exactMatch| DELIB_LF[delib:Framework] LI -->|skos:exactMatch| DELIB_LI[delib:LegalInterpretation] del -->|extends| DELIB classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#ddf,stroke:#333,stroke-width:2px

2) Argument Interchange Format (AIF) Alignment

graph TD DEL[Deliberation Ontology]:::mainNode AIF[AIF Ontology]:::ontNode A[Argument] C[Contribution] POSITION[Position] THREAD[Thread] A -->|skos:closeMatch| AIF_A[aif:Argument] C -->|skos:closeMatch| AIF_S[aif:Statement] POSITION -->|skos:closeMatch| AIF_C[aif:Claim] A -->|skos:relatedMatch| AIF_SS[aif:SupportScheme] A -->|skos:relatedMatch| AIF_AS[aif:AttackScheme] THREAD -->|skos:relatedMatch| AIF_N[aif:ArgumentNetwork] del -->|extends| AIF classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#dfd,stroke:#333,stroke-width:2px

3) SIOC Ontology Alignment

graph TD DEL[Deliberation Ontology]:::mainNode SIOC[SIOC Ontology]:::ontNode C[Contribution] DP[DeliberationProcess] P[Participant] THREAD[Thread] T[Topic] ORG[Organization] GROUP[Group] C -->|skos:broadMatch| SIOC_P[sioc:Post] DP -->|skos:relatedMatch| SIOC_F[sioc:Forum] P -->|skos:relatedMatch| SIOC_U[sioc:UserAccount] THREAD -->|skos:closeMatch| SIOC_T[sioc:Thread] T -->|skos:relatedMatch| SIOC_TO[sioc:Topic] ORG -->|skos:relatedMatch| SIOC_CO[sioc:Community] GROUP -->|skos:relatedMatch| SIOC_G[sioc:Usergroup] del -->|extends| SIOC classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#fdf,stroke:#333,stroke-width:2px

4) FOAF Ontology Alignment

graph TD DEL[Deliberation Ontology]:::mainNode FOAF[FOAF Ontology]:::ontNode P[Participant] ORG[Organization] GROUP[Group] P -->|skos:narrowMatch| FOAF_P[foaf:Person] ORG -->|skos:exactMatch| FOAF_O[foaf:Organization] GROUP -->|skos:closeMatch| FOAF_G[foaf:Group] P -->|skos:relatedMatch| FOAF_A[foaf:Agent] del -->|extends| FOAF classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#dff,stroke:#333,stroke-width:2px

5) Dublin Core (DC) Alignment

graph TD DEL[Deliberation Ontology]:::mainNode DC[DC Ontology]:::ontNode IR[InformationResource] TIMELINE[Timeline] T[Topic] IR -->|skos:broadMatch| DC_R[dc:Resource] IR -->|title property: skos:exactMatch| DC_T[dc:title] IR -->|description property: skos:exactMatch| DC_D[dc:description] TIMELINE -->|date properties: skos:relatedMatch| DC_DA[dc:date] IR -->|creator property: skos:exactMatch| DC_C[dc:creator] T -->|skos:relatedMatch| DC_S[dc:subject] del -->|extends| DC classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#ffd,stroke:#333,stroke-width:2px

6) LKIF Ontology Alignment

graph TD DEL[Deliberation Ontology]:::mainNode LKIF[LKIF Ontology]:::ontNode LS[LegalSource] LD[LegalDecomposition] LI[LegalInterpretation] LF[LegalFramework] POSITION[Position] LS -->|skos:closeMatch| LKIF_LS[lkif:LegalSource] LD -->|skos:relatedMatch| LKIF_D[lkif:LegalDocument] LI -->|skos:relatedMatch| LKIF_N[lkif:Norm] POSITION -->|when normative: skos:relatedMatch| LKIF_S[lkif:Statement] LF -->|skos:relatedMatch| LKIF_LT[lkif:LegalTheory] del -->|extends| LKIF classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#fdd,stroke:#333,stroke-width:2px

7) IBIS Model Alignment

graph TD DEL[Deliberation Ontology]:::mainNode IBIS[IBIS Model]:::ontNode T[Topic] POSITION[Position] A[Argument] DECISION[Decision] DP[DeliberationProcess] T -->|skos:closeMatch| IBIS_I[ibis:Issue] POSITION -->|skos:exactMatch| IBIS_P[ibis:Position] A -->|skos:closeMatch| IBIS_A[ibis:Argument] DECISION -->|skos:relatedMatch| IBIS_R[ibis:Resolution] DP -->|skos:broadMatch| IBIS_D[ibis:Deliberation] del -->|extends| IBIS classDef mainNode fill:#ff9,stroke:#333,stroke-width:3px classDef ontNode fill:#ddf,stroke:#333,stroke-width:2px

Each subgraph above focuses on how the Deliberation Ontology Graph maps to or extends the respective ontology's classes and properties. By splitting the overall mapping into smaller diagrams, we ensure that each relationship set is easier to inspect, maintain, and update.

Documentation

Comprehensive documentation for the Deliberation Knowledge Graph project: