Skip to content
Closer - The home of longitudinal research

European DDI (EDDI) User Conference 2023

EDDI2023 will take place in Ljubljana, Slovenia from 27-29 November 2023.

About EDDI

EDDI is the annual conference for users of DDI – Data Documentation Initiative -, a metadata specification for the social, economic, and behavioral sciences. It is run by GESIS and the IDSC of IZA under the auspices of the DDI Alliance.

EDDI is designed to provide forum where DDI users from Europe and the world can gather to showcase their work and their progress toward DDI adoption, as well as discuss any questions or challenges they may have about the standard.

EDDI includes presentations, poster sessions, and discussion sessions. The conference closes with a “meet the experts” session in which users will have a chance to lobby for their point of view with representatives from the Technical Committee of the DDI Alliance. The philosophy of EDDI is to be an open, inclusive DDI community-building activity.

About the 2023 conference

The 15th Annual European DDI User Conference (EDDI 2023) will be hosted by the Slovenian Social Science Data Archives (ADP).

ADP was established in 1997 as an organisational unit of the Social Sciences Research Institute at the Faculty of Social Sciences, University of Ljubljana. It is a national research infrastructure for social sciences, whose main mission is to manage data and data services in order to support research and education.

EDDI 2023 will be held as an in person event in Ljubljana, Slovenia with online participation.

Programme overview:

  • Tutorials and Workshops: Monday 27 November 2023
  • Conference: Tuesday 28 November – Wednesday 29 November 2023
  • Side Meetings: Thursday 30 November – Friday 1 December 2023

CLOSER’s presence at EDDI2023

The CLOSER team will run a workshop and present two papers and two posters at this year’s EDDI conference. Please see below for further details.

Workshop:

Date: Monday 27th November
Title: Metadata Uplift and Machine Learning – European Perspectives
Authors: Jon Johnson, Alina Danciu, Christophe Dzikowski, Claus-Peter Klas, András Micsik, Knut Wenzig, (Arne Bethmann)

The background for this workshop are recent calls for proposals where participants had limited understanding of the scope of metadata holdings, local computer science expertise and knowledge of current work being done at institutions to be able to develop a focused and convincing proposal for funding.

The development of the European Question Bank, and the European Language Social Science Thesaurus (ELSST), in addition to other national and cross national metadata resources could be utilised to develop methods and training data in a collaborative way that support open science.

The purpose of this workshop is to bring together metadata content providers and infrastructures that are interested in co-ordinating such efforts to uplift existing and future metadata holdings with the aim to put together future research proposals.

Expected topics for the workshop will be, available metadata, in DDI (or otherwise), current and future plans for ML or AI activities, development of cross-language training datasets, and the types of collaborative project which are possible.

Presentations:

  • Cross-Country Collaboration Efforts with Common Conceptual Groups – NACDA and CLOSER
    Kathryn Lavender, Hayley Mills, Jon Johnson, Jim McNally, Sanda Ionescu, Jennifer Zeiger
    The National Archive of Computerized Data on Aging (NACDA) and CLOSER – the interdisciplinary partnership of leading UK social and biomedical longitudinal population studies (LPS),  have been discussing ways to collaborate and create common conceptual groups across their social science data collections in their Colectica Portals.Although NACDA and CLOSER have approached metadata organization efforts in different ways, and present the metadata differently in their portals, the use of common DDI-Lifecycle metadata structures simplifies the possibilities for collaboration, in particular the mapping of overlapping concepts which have been identified across both LPS data collections.

    We plan to discuss how we have managed our collaborations thus far and include highlights from an in-person workshop we held together earlier this year. Further, we will explain some of the details of our respective approaches to creating conceptual variable groups and why these are important to our projects and to future reuse.

    NACDA is part of the Inter-university Consortium for Political and Social Research (ICPSR) and based at the Institute for Social Research (ISR) at the University of Michigan. CLOSER is based at the University College London (UCL) Social Research Institute (SRI).

  • Managing repeated representation of variables in DDI Lifecycle
    Christophe Dzikowski, Jon Johnson, Hayley Mills, Frank Cotton, Sophiane Kab
    DDI-Lifecycle utilises the variable cascade to organise and describe data from conception to collection.

    The organisation of conceptual variables and conceptual variable groups allows comparison of data at different time points, universes, representations and many other dimensions through concordance tables and is well suited to iterative data collected as panels, cohorts, repeated surveys and periodic administrative files.

    Within a single study design, such comparisons are straight-forward to describe and comprehend. Whilst it is possible to extend the comparison between studies or from other sources such as administrative data, by creating groups of conceptual variables, conveying to the user (human or machine) the data collection dimensions will likely require the alignment of this metadata also.

    This is potentially a significant challenge for multi-study metadata infrastructures or national statistical agencies which hold data from diverse sources.

    The presentation will discuss the outputs from a workshop organised between CLOSER, INSEE and Constance that seeks to address these issues.

  • Semi-automating questionnaire metadata entry for increased job satisfaction
    Becky Oldroyd, Hayley Mills, & Jenny Li
    CLOSER Discovery is the UK’s most comprehensive research tool for longitudinal population studies, containing questionnaire and dataset metadata for 11 leading UK studies.

    Creating questionnaire metadata can be a time-consuming and challenging task. Historically, CLOSER’s Metadata Assistants (MAs) entered the questionnaire metadata into our in-house developed DDI questionnaire editor – Archivist – by manually entering them into the tool.

    CLOSER are committed to creating enriching and fulfilling jobs, particularly for those who create the content that enables CLOSER Discovery to be an evolving and valuable resource. Subsequently, CLOSER’s MA role has advanced from manual metadata entry to semi-automated metadata editing using GitLab parsers.

    Gitlab is freely available for educational institutes and open-source software projects, and allows the automation of tasks through a simple interface.

    These parsers use the available structured information from the studies (e.g., PDF, XML) so that questionnaire metadata can be loaded into Archivist, and then checked and edited. Consequently, our workflow is more efficient with reduced human error and, importantly, the MA role is more fulfilling and allows staff to focus on the aspects that are most engaging and creative.

    This presentation will provide an overview of CLOSER’s GitLab parsers, and explain how they have advanced CLOSER’s MA role.

Posters:

  • CLOSER Discovery update: user-driven redesign
    Hayley Mills
    CLOSER Discovery is the UK’s most comprehensive research tool for longitudinal population studies (LPS). DDI-Lifecycle is used to document questionnaire, question, dataset and variable metadata from 11 (and counting) leading UK studies. CLOSER Discovery is powered by the Colectica software stack including a customised Colectica Portal which sits on top of Colectica Repository, and provides access to item level metadata through display as web pages. 

    Since the launch in 2017, the CLOSER Discovery portal has had only minor visual updates within the bounds of the underlying software. In the latest update, launched in September 2023, we have worked with Colectica to update the Portal software, enabling the implementation of the design created by Bravand, a digital design and build company with experience of complex websites. 

    CLOSER Discovery has been redesigned using detailed feedback from one-to-one interviews with users, focussing on the home, search and explore pages. The poster will detail the new design features and the feedback it is based on.  

  • Python Library for Colectica Portal API
    Jenny Li and Jon Johnson
    Colectica Portal is widely used to disseminate DDI-LIfecycle metadata. It acts as a website which sits on top of Colectica Repository and provides access to item level metadata through display as web pages and as downloads in a number of formats.The API is documented in the swagger documentation that is supplied with Colectica Portal, but for those unfamiliar with either API programming or DDI-Lifecycle, using the API can be a significant barrier.

    The colectica-api wrapper allows a user with little experience of using APIs or DDI-Lifecycle a straight-forward entry point to those with python programming experience to access and discovery DDI-LIfecycle metadata. The colectica-api wrapper is open sourced and available as a pip install package for ease of use, and provides example code to get a python developer up and running. The wrapper takes advantage of the introduction of JSON format in the recent versions of Colectica which is preferred over the previous XML format.

Further information

Please visit the EDDI conference website for further information including the provisional programme and details on how to register.