SciVal Metric: Awarded Grants Value

   Back to all metrics

Awarded Grants Value is the total amount of funding that has been awarded. The value of a Grant is attributed to its start year.

The metric includes Awarded Grants from Funding Bodies covered by SciVal, awarded to Institutions known in SciVal. Up-to-date coverage can be found via the “Funding Bodies covered” link at the top right corner on Grants module pages.

Awarded Grants Value is displayed in US Dollar. The US Dollar values are calculated using the historical exchange rates provided by the US Federal Reserve.

Allocation to Institution(s)

Grants are assigned to Institutions by looking at the Scopus affiliation(s) of the award investigator(s) mentioned in the Awarded Grant, at the time of it being awarded. Institutions in SciVal are made up of multiple Scopus affiliations and the affiliation is then used to set the Institution.

Based on the available investigator and affiliation data an Awarded Grant can be assigned to multiple Institutions. Exact allocation of Awarded Grants Value to each Institution is used in case that data is available.

In case we don’t have detailed information about the allocation, the Awarded Grants Value is evenly distributed to each Institution. Example: if an Awarded Grant worth $3,000,000 is allocated to six Institutions, each Institution will show an Awarded Grants Value of $500,000

Mapping to Topics

Grants can be linked to Topics through two primary methods: linked publications and semantic linking. For Grants with available linked publications, SciVal Topics of these publications are used to determine the research areas funded by the Grant.

The majority of Grants are linked to a small number of Topics, with about 90% associated with up to 10 Topics, and most of these linked to fewer than 5. However, some larger Grants have publications linked to over 10, and occasionally more than 1000 Topics; for these, the top 10 Topics with the most linked publications are chosen.

For Grants without linked publications, an in-house developed Simple Text Embedding Model (STEM) is used for assigning Topics. STEM, a Neural Network Classifier, processes text from 60 million articles on Scopus to learn which Topics they belong to, creating a semantic link between a Grant and its Topics. This method ensures Topics are assigned based on an optimal combination of precision and recall, with a limit of 10 Topics per Grant. The method used for linking each Grant to Topics is visible in the Grant preview window, along with the associated Topics.

Grants can be linked to more than one Topic, reflecting the often broad nature of research Grants compared to the more granular SciVal Topics. When a Grant is linked to multiple Topics, the full awarded Grant value is assigned to each Topic but is deduplicated at the Topic Cluster level.

The assignment of Grants to Topics is reviewed annually in June to incorporate any new linked publications, ensuring the Topic associations remain current and accurate.

This metric is useful to…

• Analyse funding trends in Topic, Topic Cluster or Topic Group with funding received from various Funding bodies covered in Grants.
• Get a sense of funding from a Funding body overall.
• Assess how different Topics, Topic Clusters or Topic Groups compare in their funding.