DemocraSci - A Parliamentary Knowledge Graph (4 legislative periods)
DOI10.5281/zenodo.13920293Zenodo13920293MaRDI QIDQ6688304FDOQ6688304
Dataset published at Zenodo repository.
Lilian Gasser, Fernando Perez-Cruz, Kourosh Shariat, Vincent Jung, Luis Salamanca, Anna Schmidt-Rohr, Laurence Brandenberger, Frank Schweitzer, Julian Minder, Sophia Schlosser, Leon Babić, Marta Balode
Publication date: 15 October 2024
Copyright license: Creative Commons Attribution 4.0 International
Overview The current dataset encodes policy-making and parliamentary processes in a knowledge graph (KG). This structure allows for the detailed representation of relationships between various entities and tracks the evolution of bills from their initial proposal through to the final vote. It also includes comprehensive metadata on the members of parliament (MPs), such as those who proposed the bills and participated in discussions across different chambers. Additionally, the dataset contains the full text of speeches delivered by MPs during these deliberations. Files contained The DemocraSci dataset comprises the following files: neo4j.dump: the file containing the Neo4J graph database. democrasci_rdb.db :a SQL database formed by the speeches given by MPs in the Parliament chambers. The DB contains a single table with two columns, one with the speech transcript and another with a unique identifier. This uid links each speech with a Speech node in the KG. The previous information is extracted for four consecutive legislative periods (48 to 51), from year 2007 to 2023. The structure of this KG is detailed in https://democrasci.jkminder.ch/index.html Here, we provide more information on the different nodes and relations contained in the graph, as well as the properties of each. Dataset creation The dataset has been parsed from the CuriaDB database of the Swiss parliament. From the information contained within, we defined a schema that encodes the political process and all the entities involved in it. All the different extraction and processing tasks, from the parsing of the original database to the curation of the data and the manipulation of the KG are done using Python. Specifically, for the population of the Neo4J database we leveragedData2Neo. Resources The knowledge graph has been created using Neo4J. The Community Edition is free to use and already offers many tools for KG exploration and graph data science https://neo4j.com/deployment-center/#community No4J can be used through the command line interface, launching a web instance where previously a Neo4J dump has been loaded. https://neo4j.com/docs/operations-manual/current/backup-restore/restore-dump/ Additionally, the Neo4J desktop app facilitates restoring a dump, and enables using more tools for exploration, such as Bloom https://neo4j.com/docs/desktop-manual/current/operations/create-from-dump/ For the exploration of the SQL DB, the easiest option is DB Browser
This page was built for dataset: DemocraSci - A Parliamentary Knowledge Graph (4 legislative periods)
