Using GRID to discover and analyse global research trends

New Digital Research Report by Martin Szomszor and Jonathan Adams highlights the need for a reference database of research organisations

(2 March 2017) The latest Digital Research Report, Discovery and Analysis of Global Research Trends Using GRID, is published today by Digital Science. Using data from the open access journal PLOS ONE, the authors highlight the need for a reference database of research organisations through analysis of author affiliation strings on articles published between 2006-2016. The report demonstrates the utility of our Global Research Identifier Database (GRID), both as a platform for data consolidation, and as an essential analytical component supplying various opportunities for geographic and sector-based analyses.

As research information management becomes more sophisticated and the diversity of research data available increases, the need to address the issue of “location” becomes ever more important – being able to accurately and unambiguously associate research data to the appropriate organisations, regions, or countries is an essential step in any analysis.

This is a particularly hard task on two fronts: i) the data referencing organisations is often unstructured, hugely varied, and contains ambiguous references to organisations; ii) the domain being modelled (namely the global network of research organisations) is complex, changing over time, and varies in structure between countries.

GRID addresses these two issues using a combination of automatic and manual techniques. The database itself has been hand crafted by a team of curators, following a set of policies put in place that describe a tractable and globally consistent view of research organisations. Our disambiguation algorithms are based on a combination of geographic entity recognition and a large database of manually curated name variants from real data sources.

The report released today showcases the capability and utility of GRID for the purposes of collaboration analysis in journal articles, but is equally useful when processing a range of other research data including grants, patents, books, conference proceedings, clinical trials, and many more.

The report:

  • Presents an analysis of how author affiliation strings vary, using 1.4M examples from a 10 year period
  • Shows how GRID can be used to measure, compare and track collaborations patterns
  • Includes a visualisation of the global collaboration network, highlighting the activity in key research institutions
  • Maps the world’s research intensity at a regional level, and compares US State collaboration patterns to those in Europe and Asia
  • Identifies healthcare collaboration precincts where Universities and local hospitals      collaborate intensively

GRID is freely available under a CC0 license in a variety of formats (CSV, JSON, RDF) and can be downloaded from grid.ac/downloads.

Download the report here.

The announcement in full is here.