Universities and research institutions must understand not only where their research strengths lie, but how those strengths are evolving over time. Identifying emerging areas of research now plays a central role in shaping institutional strategy, informing investment and hiring decisions, and responding to funder priorities and wider societal challenges.
Imagine an ever-expanding body of scholarly knowledge: dense, interconnected, and increasingly complex. As research becomes more specialised, it also spans a growing range of disciplines. Ideas, methods, and concepts increasingly cross disciplinary boundaries and are often difficult to disentangle. The result is a research landscape that is rich in insight, but inherently noisy.
How can universities begin to address these challenges in practice? The methodological work outlined by Hélène Draux in From noise to signal: how to find emerging research trends in Dimensions explores this question directly, examining how large-scale publication data can be interpreted more effectively to identify emerging research activity.
Older approaches are falling short
Institutions commonly rely on a combination of established approaches to identify research trends. These include existing classification systems, author-supplied keywords, curated vocabularies, and more manual forms of horizon scanning. While widely used, each of these approaches has limitations when applied to large, evolving, and multidisciplinary research landscapes.
Classification systems are often relatively shallow or lag behind emerging areas of research. Author-supplied keywords can be inconsistent, strategically chosen, or highly variable across disciplines. Curated vocabularies such as MeSH offer high-quality control, but are typically top-down and slow to reflect new or rapidly evolving research directions.
More data-driven approaches, such as machine learning-based concept extraction, can address some of these limitations, but often require significant technical effort, including large-scale data acquisition and processing. As a result, institutions may struggle to apply such methods consistently or at scale.
Differences in publication practices further complicate interpretation: what appears to be an emerging trend in one field may look very different in another.
From noise to signal: a methodological perspective, applied in Dimensions
Draux’s work highlights that research concepts extracted automatically from large publication datasets are inherently noisy. Without filtering, analyses risk being overwhelmed by terms that are frequent but analytically uninformative. This is where tools such as Dimensions can play an important role.
Dimensions provides the analytical context in which this methodological perspective can be applied at scale. Research concepts are extracted across publications, with relevance scores indicating conceptual centrality. Longitudinal data supports exploration of how concepts evolve over time and across disciplines.
This enables exploratory analysis that can be adapted to different institutional and disciplinary contexts. Quantitative signals are intended to support interpretation rather than replace expert judgement, complementing qualitative insight and peer review.
Implications for academic strategy
For academic institutions, this approach supports a more nuanced understanding of research activity. It helps distinguish between established and emerging areas, identify potential future strengths earlier, and ground strategic discussions in evidence that reflects both relevance and change over time.
Used carefully, such analyses can inform conversations about focus, investment, and development, while recognising the limits of quantitative indicators and the importance of disciplinary expertise.
Going deeper into the methodology
Draux’s article provides a high-level overview of an approach to identifying emerging research trends. Readers interested in the detailed reasoning behind concept relevance, trend categorisation, and the challenges of working with large-scale publication data are encouraged to read her original post, From noise to signal: how to find emerging research trends in Dimensions, which offers a deeper exploration of these ideas.
Stay informed about research analytics and methodology
Subscribe to the Dimensions newsletter for occasional updates on research evaluation, trend analysis, and the use of data to support academic strategy.
Discuss your institutional context
If you would like to explore how these approaches could be applied within your own institution, the Dimensions team is available for an exploratory conversation.