multiobs

Our Recent Productions

Datasets, working papers, and peer-reviewed articles produced by MultiObs researchers.

Showing 1–3 of 3

  1. Working Paper2026Open Access
    Collective action for Open Research Information: Introducing ORION-DBs and a new FORCE11 Working Group(opens in a new tab)

    Mazoni, Alysson; Kramer, Bianca; Neylon, Cameron; Costas, Rodrigo; van Eck, Nees Jan

    With ORION-DBs, a growing set of actors are hosting large research information datasets in shared spaces, notably in Google BigQuery. To coordinate amongst current and future contributors and work towards community standards for discovery, processing, preservation and documentation, we are starting a FORCE11 working group. The working group will also seek to build a community around future alternatives to proprietary cloud systems for data sharing at scale and identify future paths to community-led training and development resources. Our goal is to build a community with a focus on supporting the users of Open Research Information resources to make more sophisticated and complex use of these data sources. We will do this by consolidating and expanding the resources themselves through supporting providers, gathering user stories and developing training materials. Finally we will scope pathways to technological independence from proprietary systems consistent with the needs of users. This webinar will introduce the ORION-DB initiative, present a number of use cases and introduce the FORCE11 Working Group. If you are a data provider, metadata user or infrastructure developer and want to learn more about ORION-DB and the planned activities of the working group, you are invited to attend the webinar. The version of this slide deck on Google was originally: https://docs.google.com/presentation/d/1ua88tr2gP2HTurQPkM4-A7eWjFiDLUI_ht3MJdHxOxc/edit?usp=sharing

  2. Working Paper2026Open Access
    A global map of science according to a topic model of EU grants(opens in a new tab)

    Vidal, Adrià Plazas; Bovenzi, Nicandro; Durán-Silva, Nicolau; Martí, Enric Fuster; Ràfols, Ismael

    Since their appearance in the 1980s, global maps of science have proved useful to locate the relative position of topics and organisations in large scientific landscapes, providing strategic information for portfolio analyses in science policy. At this global level (global meaning here ‘all disciplines’), these maps have been based on publications and have mostly used citations to build the relationship between scientific fields. Here, we present a comprehensive scientific map based on grant data that uses semantic analyses (BERT-based topic modelling) to provide the overall landscape of science based on grant data from the European Union since 2003 (since the 6th framework programme). We find a disciplinary structure which is consistent with the circular shape of publication and citation based global science maps. However, this grant map shows a much larger presence of applied science and mission-oriented research, such as engineering and sustainability sciences. We provide examples of how this landscape can be used as a baseline to explore the specialisation of territories (regions), organisations (universities) and funding programmes.

  3. Working Paper2025Open Access
    Exploring Interoperability Between Local and Global Databases in Scientometrics: Lattes, Capes, and OpenAlex(opens in a new tab)

    Mazoni, Alysson Fernandes; Borges, Luís Fabiano Farias; Macedo, Estevao Fernandes; Tuesta, Esteban Fernandez

    Numerous initiatives are currently underway to disambiguate databases worldwide. In this paper, we propose a methodology for disambiguating research entities using big data techniques, adopting an approach that goes from local to global databases. Our objective is to enhance the quality of data in the OpenAlex database by leveraging information from Brazilian databases, particularly data from the Lattes Platform and the Brazilian Federal Agency for Support and Evaluation of Graduate Education. We compare similar names of authors and institutions, employing Digital Object Identifiers to link entities, along with an adaptation of the Levenshtein distance algorithm. The proposed method is straightforward to implement in tabular databases and facilitates disambiguation, thereby contributing to open science practices and providing an effective solution for research information systems. The findings indicate the potential for integrating local and global databases to address issues related to ambiguous names and incomplete metadata.