OpenDataGOV-AP: A Linked Data Application Profile for parliamentary activities

1 Introduction

Before elections, each political party presents an electoral program that outlines its commitments. Numerous researchers worldwide have examined the fulfilment of these programs and generally agree that most parties do fulfil their promises[1–5], provided they have the necessary conditions to do so[6]. Despite this academic consensus, the general public often perceives political parties negatively regarding their ability to keep these promises[4]. Serra-Silva et al.[6] argue that many studies compare electoral programs with implemented policies while overlooking legislative initiatives that were not accepted. They suggest that evaluating opposition fulfilment should rely on proposed legislative initiatives rather than just accepted ones, as this approach provides a more comprehensive and accurate estimate of each party's efforts.

The Open Government Partnership (OGP) is an organisation that aims to promote better governance by encouraging transparency, inclusiveness, accountability, and participation¹. Portugal has been an OGP member since 2017², but the Portuguese plan, according to a recent OGP report³, has problems. The Portuguese government provides numerous websites with open data, including information on legislative initiatives⁴, members of parliament⁵, public contracts⁶ and many other public datasets⁷. However, the data is scarce in many different XML documents, and much information is hidden in PDF documents. The data from the Portuguese Parliament needs to be made accessible to citizens and other stakeholders in a format that breaks the silo of the parliament. A linked dataset already exists[7]; however, it is limited to information about members of parliament, leaving broader information needs unaddressed.

Lourenço et al.[8] identify two key societal stakeholder groups. The first comprises “ordinary” citizens who, lacking formal roles in government and the technical skills required, face significant challenges in accessing and interpreting data published on open government platforms. The second group consists of “information brokers”—such as journalists, NGOs, and academic researchers—who possess the expertise and resources needed to effectively navigate and analyse this data.

While information brokers can rely on established strategies to remain informed, ordinary citizens are often left at a disadvantage. Limited time, insufficient technical knowledge, and restricted access to analytical tools hinder their ability to engage critically with available information. This imbalance creates barriers to understanding and increases the risk of misconceptions. Ideally, such a gap should not persist; however, the lack of accessible tools continues to reinforce it, granting a disproportionate advantage to those already equipped with the necessary skills.

As a result, the promise of transparency in open government data remains only partially fulfilled, benefiting a relatively narrow segment of society. This asymmetry becomes particularly problematic in contexts where misinformation and fake news are used to influence democratic processes. Citizens who are unable to independently verify or contextualise information are more vulnerable to manipulation, contributing to a self-reinforcing cycle of distrust and misinformation. Addressing this issue requires the urgent development of inclusive and accessible tools that empower a broader range of users to engage with public data, thereby strengthening informed participation and public trust.

This paper addresses this problem by presenting a linked data application profile, which is a schema to publish linked data in RDF about Portuguese Parliament Initiatives and related actors, relationships between legislative documents, organisations, and policy topics.

This paper is organised as follows. Section 2 provides background on Information Retrieval, semantic search, knowledge graphs, and application profiles, establishing the theoretical context for the work. Section 3 describes the methodology, detailing the data-driven development of the data model across four modules — Core, Initiatives, Biographical-Profile, and MP-Activities — followed by the construction of the constraints matrix and the vocabulary alignment process. Section 4 presents the results, comprising the OpenDataGOV-AP itself and its associated SHACL file, the implementation of the Knowledge Graph through a Python-based parser and RDF converter, and an evaluation through a SPARQL query that demonstrates the profile's applicability and future impact on the community that will serve. Section 6 concludes the paper with remarks on the contribution and directions for future work.

2 Background

3 Methodology

The development of the OpenGOV Linked Data Application Profile (OpenGOV-AP) was informed by previous work[27–29]. In this context, we followed the deployment guidelines proposed by Tim Berners-Lee[30], aiming to transform the three-star data in the XML files and one-star data in the PDF files into five-star data. Additionally, our design was guided by the FAIR principles¹⁰, which emphasise Findability, Accessibility, Interoperability, and Reusability to promote effective management and long-term usability of the data.

During development, we began by defining the data model. Next, we established the constraints matrix, defining for each term its domain and range, its cardinality and any encoding scheme (value or syntax) for the values, to reduce ambiguity and enhance downstream retrieval. Finally, we proceeded with vocabulary alignment, ultimately selecting the most appropriate RDF vocabulary terms. The classes and terms were aligned with ontology-engineering principles to make the semantics explicit and machine-processable[18–20].

The following sections present the development process and its milestones, including the techniques and methods used.

4 Results, Implementation and Validation

5 Discussion on Impact of the work

The results on Table B1 reveal cross-party collaboration patterns that aren't visible from voting records alone. Specifically:

• For journalists:

  • Pairs with high co-authorship despite opposing parties signal pragmatic alliances worth investigating — is there a deal, shared constituency interest, or personal relationship driving it?

  • Common ideologies surfaced from Wikidata can expose when parties market themselves as distinct but share underlying political DNA — useful for "parties are more alike than they admit" stories.

  • Outliers (e.g., far-left and far-right MPs co-authoring) are inherently newsworthy.

• For policy-makers:

  • High co-authorship pairs identify natural coalition-building opportunities for getting legislation passed.

  • If certain ideological overlaps consistently produce more collaborative output, that's a signal for which parties to negotiate with on specific issue areas.

  • The exclusion of PSD&CDS (the parties at the moment in the Government) means the analysis focuses on opposition dynamics — useful for understanding how minority parties coordinate.

The results on Table B2 show that the combination of all five metrics in one table is what makes it powerful — each metric alone is interesting, but together we can ask whether underrepresented constituencies are also underserved legislatively. The data connects representational input (voters, seats) with legislative output (initiatives), which opens several angles:

• For journalists:

  • Constituencies where MPs submit far fewer initiatives per 100,000 voters than average are underperforming — a straightforward accountability story.

  • The "voters per seat" metric exposes malapportionment: some voters effectively have more representation than others, which is a structural fairness story.

  • Comparing high-electorate vs. low-electorate constituencies reveals whether smaller districts punch above their weight legislatively.

• For policy makers:

  • Persistent low initiative output from certain constituencies may indicate MPs are overloaded, under-resourced, or disengaged — a case for institutional support or electoral reform.

  • The voters-per-seat ratio feeds directly into debates about redrawing constituency boundaries or reforming the electoral system.

  • Normalising by 100,000 voters makes it defensible as a KPI for representational equity across very differently sized districts.

• For the general public:

  • People can look up their own constituency and see how active their MPs are relative to others — a rare, concrete accountability tool.

  • The voters-per-seat figure makes abstract concepts like "vote weight" tangible: your vote may literally count less than someone else's depending on where you live.

  • It demystifies parliament by translating legislative activity into a familiar unit (per 100,000 people), similar to how health or crime statistics are typically presented.

When legislative data is open, semantised, and machine-readable, it becomes more than a record — it becomes a tool. Anyone can query it, connect it to other knowledge sources, and tell accurate, evidence-based stories about how power is exercised. That kind of radical transparency is one of democracy's best defences against misinformation and populism.

6 Conclusion and Future Work

This paper addressed the challenge of making Portuguese parliamentary data more accessible and reusable through a Linked Data Application Profile (OpenDataGOV-AP). Starting from heterogeneous XML and PDF sources, we defined a semantic model (a Linked Data Application Profile) that structures parliamentary initiatives, actors, institutional relations, and policy-related metadata within an RDF framework.

The resulting Linked Data Application Profile establishes a foundation for publishing and integrating parliamentary data beyond isolated systems, and for supporting downstream applications such as semantic search. By explicitly modelling entities, properties, and relationships relevant to legislative activity, this work helps reduce technical barriers to accessing public information and strengthen transparency-oriented data infrastructures. modelling As future work, we plan to complete and validate the full profile implementation with real publishing pipelines, enrich the model with additional parliamentary and electoral sources, and evaluate retrieval effectiveness in a semantic search prototype.

Acknowledgments

This work is funded by the Universidade do Porto - Faculdade de Engenharia.

During the preparation of this manuscript, the authors used Grammarly.com to improve the readability, grammar, and overall structure of the text. The authors reviewed, edited, and validated the resulting content and take full responsibility for the final version of the manuscript.

References

1. [1] A. M. Belchior, C. Moury, L. Hanenberg, and P. Nunes, O cumprimento dos programas eleitorais em Portugal: Tendências, reformas políticas e o papel dos veto players (1995-2019). in Análise Social, pp. 160-184, 2022. https://scielo.pt/scielo.php?script=sci_arttext&pid=S0003-25732022000100160&nrm=iso.
2. [2] J. Artés, Do spanish politicians keep their promises?. in Party Politics, vol. 19, pp. 143-158, 2013.
3. [3] E. Naurin, Is a promise a promise? Election pledge fulfilment in comparative perspective using sweden as an example. in West European Politics, vol. 37, pp. 1046-1064, 2014. https://doi.org/10.1080/01402382.2013.863518.
4. [4] C. Moury, Minority governments and pledge fulfilment: Evidence from portugal. in Government and Opposition, 2016. https://doi.org/10.1017/gov.2016.14.
5. [5] R. Thomson, Citizens' evaluations of the fulfilment of election pledges: Evidence from ireland. in The Journal of Politics, vol. 73, 2011. https://doi.org/10.1017/S0022381610000952.
6. [6] S. Serra-Silva and A. M. Belchior, Understanding the pledge fulfilment of opposition parties using evidence from portugal. in European Politics and Society, vol. 21, pp. 72-90, 2020. https://doi.org/10.1080/23745118.2019.1605961.
7. [7] L. Inês, B. Ana Alice, and A. Óscar Atanázio, Map4Scrutiny – A Linked Open Data Solution for Politicians Interest Registers. in Proceedings of the International Conference on Dublin Core and Metadata Applications, Dublin Core Metadata Initiative, 2023. https://doi.org/10.23106/DCMI.953176634.
8. [8] R. P. Lourenço, Data disclosure and transparency for accountability: A strategy and case analysis. in Information Polity, vol. 18, pp. 243-260, 2013. https://doi.org/10.3233/IP-130310.
9. [9] R. Chauhan, R. Goudar, R. Sharma, and A. Chauhan, Domain ontology based semantic search for efficient information retrieval through automatic query expansion. in 2013 International Conference on Intelligent Systems and Signal Processing (ISSP), pp. 397-402, 2013. https://doi.org/10.1109/ISSP.2013.6526942.
10. [10] S. Neji, T. Chenaina, A. M. Shoeb, and L. B. Ayed, SemApp: A Semantic Approach to Enhance Information Retrieval. in Computational Science and Its Applications – ICCSA 2021, Springer International Publishing, Cham, 2021.
11. [11] M. Tang, J. Chen, and H. Chen, SemOIR: An ontology-based semantic information retrieval system. in 2020 IEEE 20th International Conference on Software Quality, Reliability and Security Companion (QRS-C), pp. 204-208, 2020. https://doi.org/10.1109/QRS-C51114.2020.00042.
12. [12] S. Redekar, V. Chekkala, S. Gouda, S. Yalgude, B. E. Student, B. E. Student, and B. E. Student, Web Search Engine Using Ontology Learning. in International Journal of Innovative Research in Computer and Communication Engineering, vol. 5, 2017.
13. [13] G. D and A. Rastogi, Integrating Web 3.0 Technologies for Advanced Information Retrieval Systems in the Educational Sector. in 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI), pp. 200-204, 2023. https://doi.org/10.1109/ICCSAI59793.2023.10421149.
14. [14] S. Maya, GDRMA: Graph Neural Networks for Document Retrievals With Mean Aggregation. in IEEE Access, vol. 12, pp. 185706-185727, 2024. https://doi.org/10.1109/ACCESS.2024.3512597.
15. [15] S. Jiang, T. F. Hagelien, M. Natvig, and J. Li, Ontology-based semantic search for open government data. in 2019 IEEE 13th International Conference on Semantic Computing (ICSC), pp. 7-15, 2019. https://doi.org/10.1109/ICOSC.2019.8665522.
16. [16] A. M. Fedotov, J. A. Tusupov, M. A. Sambetbayeva, S. K. Sagnayeva, A. A. Bapanov, A. N. Nurgulzhanova, and A. S. Yerimbetova, Using the thesaurus to develop it inquiry systems. in Journal of Theoretical and Applied Information Technology, vol. 86, 2016. https://www.jatit.org/volumes/Vol86No1/6Vol86No1.pdf.
17. [17] S. Sharma, A. Kumar, and V. Rana, Ontology Based Informational Retrieval System on the Semantic Web: Semantic Web Mining. in 2017 International Conference on Next Generation Computing and Information Systems (ICNGCIS), pp. 35-37, 2017. https://doi.org/10.1109/ICNGCIS.2017.21.
18. [18] T. R. Gruber, A translation approach to portable ontology specifications. in Knowledge Acquisition, vol. 5, pp. 199-220, 1993. https://doi.org/10.1006/knac.1993.1008.
19. [19] R. Studer, V. R. Benjamins, and D. Fensel, Knowledge engineering: Principles and methods. in Data & Knowledge Engineering, vol. 25, pp. 161-197, 1998. https://doi.org/10.1016/S0169-023X(97)00056-6.
20. [20] K. I. Kotis, G. A. Vouros, and D. Spiliotopoulos, Ontology engineering methodologies for the evolution of living and reused ontologies: Status, trends, findings and recommendations. in The Knowledge Engineering Review, vol. 35, 2020. https://doi.org/10.1017/S0269888920000065.
21. [21] L.-D. Ibáñez, E. Simperl, R. Palma, J. Domingue, and P. Casanovas, Crossing the chasm between ontology engineering and application development: A survey. in Journal of Web Semantics, vol. 70, 2021. https://doi.org/10.1016/j.websem.2021.100655.
22. [22] The Description Logic Handbook: Theory, Implementation, and Applications. Cambridge University Press, 2003.
23. [23] I. Horrocks, P. F. Patel-Schneider, and F. van Harmelen, From SHIQ and RDF to OWL: The making of a web ontology language. in Journal of Web Semantics, vol. 1, pp. 7-26, 2003. https://doi.org/10.1016/S1570-8268(03)00030-8.
24. [24] S. T. Sukumar, C.-H. Lung, and M. Zaman, Knowledge Graph Generation for Unstructured Data Using Data Processing Pipeline. in 2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC), pp. 466-471, 2023. https://doi.org/10.1109/COMPSAC57700.2023.00068.
25. [25] A. Oarga, M. Hart, A. M. Bran, M. Lederbauer, and P. Schwaller, Scientific knowledge graph and ontology generation using open large language models. in AI for Accelerated Materials Design-NeurIPS 2024, 2024. https://openreview.net/forum?id=wMMhffCxXZ.
26. [26] R. Heery, M. Patel, and et al., Application profiles: mixing and matching metadata schemas. in Ariadne, vol. 25, 2000.
27. [27] M. C. Malta, Contributo metodológico para o desenvolvimento de perfis de aplicação no contexto da Web Semântica. 2014. Ph.D. thesis
28. [28] M. C. Malta and A. A. Baptista, A method for the development of dublin core application profiles (me4dcap v0. 1): a description. in Proceedings of the International Conference on Dublin Core and Metadata Applications, Dublin Core Metadata Initiative, 2013.
29. [29] M. Curado Malta, P. Centenera, and E. González-Blanco, Using Reverse Engineering to Define a Domain Model: The Case of the Development of a Metadata Application Profile for European Poetry. in Developing Metadata Application Profiles, IGI Global, pp. 146-180, 2017. https://doi.org/10.4018/978-1-5225-2221-8.ch007.
30. [30] T. Berners-Lee, Linked data - design issues. 2009. https://www.w3.org/DesignIssues/LinkedData.html.

Appendix A Parliament-related ontologies

The "Archived" LEX ontology refers to https://interoperable-europe.ec.europa.eu/collection/eu-semantic-interoperability-catalogue/solution/cen-metalex/distribution/metalex-owl-ontology -- accessed in April 2026. The "404 error" Popolo entry refers to https://www.w3.org/ns/opengov -- accessed in April 2026.

Appendix B SPARQL Queries and Query Results

This Appendix presents the queries to address the questions in Section 4.3.

PREFIX m4s: <http://purl.org/map4scrutiny-core#>
PREFIX polis-ini: <http://purl.org/polis/ar/initiatives#>
PREFIX polis-core: <http://purl.org/polis/ar/core#>
PREFIX polis-graph: <http://purl.org/polis/ar/graph#>
PREFIX ocd: <http://dati.camera.it/ocd/>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
PREFIX qudt: <http://qudt.org/schema/qudt/>
PREFIX dbo: <http://dbpedia.org/ontology/>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>

SELECT ?legislatureNumber ?mp1ParliamentaryName ?mp2ParliamentaryName ?party1Acronym ?party2Acronym (COALESCE(GROUP_CONCAT(DISTINCT ?commonIdeology; separator=", "), "") AS ?commonIdeologies) ?coauthoredInitiatives
WHERE {
  {
    SELECT ?legislatureNumber ?mp1ParliamentaryName ?mp2ParliamentaryName ?party1Acronym ?party2Acronym ?party1 ?party2 (COUNT(DISTINCT ?initiative) AS ?coauthoredInitiatives)
    WHERE {
      ?initiative polis-ini:hasAuthor ?mp1, ?mp2 ;
                  polis-core:belongsToLegislature ?legislature .
      FILTER (STR(?mp1) < STR(?mp2))

      ?mp1 a m4s:Parliamentarian ; foaf:name ?mp1Name ;
           ocd:rif_mandatoCamera ?mandate1 .
      ?mp2 a m4s:Parliamentarian ; foaf:name ?mp2Name ;
           ocd:rif_mandatoCamera ?mandate2 .

      ?mandate1 polis-core:belongsToLegislature ?legislature ;
                polis-core:parliamentaryName ?mp1ParliamentaryName ;
                polis-core:hasMembership [ polis-core:isMemberOf ?party1 ] .
      ?mandate2 polis-core:belongsToLegislature ?legislature ;
                polis-core:parliamentaryName ?mp2ParliamentaryName ;
                polis-core:hasMembership [ polis-core:isMemberOf ?party2 ] .

      ?legislature dbo:number ?legislatureNumber .
      ?party1 qudt:acronym ?party1Acronym .
      ?party2 qudt:acronym ?party2Acronym .

      FILTER (?party1 != ?party2)
      FILTER (?party1Acronym != "PSD" && ?party2Acronym != "PSD")
    }
    GROUP BY ?legislatureNumber ?mp1ParliamentaryName ?mp2ParliamentaryName ?party1Acronym ?party2Acronym ?party1 ?party2
    ORDER BY DESC(?coauthoredInitiatives)
    LIMIT 20
  }

  OPTIONAL {
    ?party1 owl:sameAs ?wd1 .
    ?party2 owl:sameAs ?wd2 .

    SERVICE <https://query.wikidata.org/sparql> {
      ?wd1 wdt:P1142 ?ideologyNode .
      ?wd2 wdt:P1142 ?ideologyNode .
      ?ideologyNode rdfs:label ?commonIdeology .
      FILTER(LANG(?commonIdeology) = "pt")
    }
  }
}
GROUP BY ?legislatureNumber ?mp1ParliamentaryName ?mp2ParliamentaryName ?party1Acronym ?party2Acronym ?party1 ?party2 ?coauthoredInitiatives
ORDER BY DESC(?coauthoredInitiatives)

PREFIX polis-core: <http://purl.org/polis/ar/core#>
PREFIX polis-graph: <http://purl.org/polis/ar/graph#>
PREFIX polis-ini: <http://purl.org/polis/ar/initiatives#>
PREFIX m4s: <http://purl.org/map4scrutiny-core#>
PREFIX ocd: <http://dati.camera.it/ocd/>
PREFIX pe: <https://ukparliament.github.io/ontologies/election/election-ontology.ttl#>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX wd: <http://www.wikidata.org/entity/>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>

SELECT ?constituencyName
  ?electorate
  ?seats
  (ROUND((?electorate / ?seats) * 10) / 10 AS ?electoratePerSeat)
  (?initiatives AS ?initiativeCount)
  (ROUND((?initiatives / ?electorate) * 100000 * 10) / 10 AS ?initiativesPer100kElectors)
WHERE {
  ?constituency a pe:ConstituencyArea ;
                foaf:name ?constituencyName ;
                owl:sameAs ?wd .

  OPTIONAL {
    SELECT ?constituency (COUNT(DISTINCT ?initiative) AS ?initiativeCountRaw)
    WHERE {
      ?initiative a polis-ini:Initiative ;
                  polis-ini:hasAuthor ?mp ;
                  polis-core:belongsToLegislature ?legislature .

      ?mp a m4s:Parliamentarian ;
          ocd:rif_mandatoCamera ?mandate .

      ?mandate polis-core:belongsToLegislature ?legislature ;
               polis-core:inConstituency ?constituency .
    }
    GROUP BY ?constituency
  }

  BIND(COALESCE(?initiativeCountRaw, 0) AS ?initiatives)

  SERVICE <https://query.wikidata.org/sparql> {
        OPTIONAL { ?wd wdt:P1831 ?electorate } .
        OPTIONAL { ?wd wdt:P1410 ?seats } .
  }
}
ORDER BY DESC(?electoratePerSeat)

Notes

1. 1.

   See https://www.opengovpartnership.org/about/ - accessed in September 2025.

   ↩
2. 2.

   See https://www.opengovpartnership.org/members/portugal/ - accessed in September 2025.

   ↩
3. 3.

   See https://www.opengovpartnership.org/documents/portugal-action-plan-review-2023-2027/ - accessed in September 2025.

   ↩
4. 4.

   See https://www.parlamento.pt/ActividadeParlamentar/Paginas/IniciativasLegislativas.aspx - accessed in October 2025.

   ↩
5. 5.

   See https://www.parlamento.pt/DeputadoGP/Paginas/Deputados_ef.aspx - accessed in October 2025.

   ↩
6. 6.

   See https://www.base.gov.pt/base4 - accessed in October 2025.

   ↩
7. 7.

   See https://dados.gov.pt/pt/ - accessed in October 2025.

   ↩
8. 8.

   A thesaurus is a dictionary organised to clarify the relationships between concepts[16].

   ↩
9. 9.

   From now on, we use the term "RDF vocabulary" to refer to either an ontology or a metadata scheme.

   ↩
10. 10.

    See https://www.go-fair.org/fair-principles/, accessed in March 2026.

    ↩
11. 11.

    See https://politrackapp.com/ontologies/OpenDataGOV-AP/DataModel/core/core.png.

    ↩
12. 12.

    See https://politrackapp.com/ontologies/OpenDataGOV-AP/DataModel/initiatives/initiatives.png.

    ↩
13. 13.

    Diário da República is the official publication where the laws are published in the Portuguese Republic.

    ↩
14. 14.

    See https://politrackapp.com/ontologies/OpenDataGOV-AP/DataModel/biographic/biographic.png.

    ↩
15. 15.

    See https://politrackapp.com/ontologies/OpenDataGOV-AP/DataModel/mp-activity/mp-activity.png.

    ↩
16. 16.

    The constraint matrix is not presented in this paper due to the dimension of the matrix.

    ↩
17. 17.

    Accessed in April 2026.

    ↩
18. 18.

    See https://github.com/politrackpt/ontologies/tree/main/POLIS/AR - accessed in June 2026.

    ↩
19. 19.

    See https://docs.python.org/3/library/dataclasses.html - Accessed in April 2026.

    ↩
20. 20.

    See https://rdflib.readthedocs.io/en/stable/index.html - Accessed in April 2026.

    ↩