Skip to content
TAKOTSUBO®Life Sciences
Let's talk
Data Science & Analytics

How Data Science Mapped 45 Years of Infrared Thermography in Health

Published research | Quality & Quantity, Springer Nature

Infrared thermography on a global scale: how data science reveals the maturation of a field

A new scientometric study brought together four and a half decades of scientific production to understand the evolution of infrared thermography in health—and shows how an auditable data architecture can turn thousands of publications into strategic knowledge.

Infrared thermography is valued for recording thermal patterns on the body surface without ionizing radiation. But how has the field grown? Which questions have come to the fore? Which countries form the strongest scientific connections? And how can impact be compared without mistaking citation volume for relative quality?

These questions guided the article Infrared thermography in biomedical and health-related research: a scientometric study based on Scopus and Web of Science (1980-2025), published in Quality & Quantity. The study was developed with the collaboration of the Research and Development Division – Data Science & Analytics at Takotsubo Life Sciences, bringing together data curation, bibliometrics, network analysis and scientific visualization.

Global map of scientific production in health-related infrared thermography, 1980-2025, with a colour scale from 1 to 1,902 articles

Total articles

Geographic distribution of scientific production in health-related infrared thermography. Each country is coloured by the total number of articles identified from 1980 to 2025. Source: Scopus and Web of Science; prepared by João Alberto de Souza Ribeiro, Instituto Termodiagnose Brasil. DOI: 10.1007/s11135-026-02890-z.
12,246documents in the analytical window
137countries identified through affiliations
3,721scientific journals mapped
1980–2025historical time span analysed

From bibliographic databases to a reliable view of the field

The starting point was the scientific record held in Scopus and Web of Science. The task was not simply to count publications, but to construct a corpus that could be inspected, reproduced and questioned transparently.

1. Integration and deduplicationRecords from the two databases were combined. Identifiers including DOI, title, publication year and authorship were used to recognise when both sources described the same paper. The result was a canonical set of 12,255 unique documents, with traceability retained between source record and publication.
2. Harmonisation and normalisationCitation data were harmonised at document level. To compare countries and research areas across different periods, the study used field- and year-normalised indicators, preventing natural differences in visibility between disciplines—or between newer and older papers—from distorting interpretation.
3. Networks and thematic structureCountries were connected through international co-authorship, and keywords through co-occurrence. Community-detection algorithms made it possible to identify both the major research programmes and the long tail of highly specialised topics.
4. Verification and reproducibilityThe R workflow was modular and deterministic, with intermediate artefacts, diagnostics and checks for coverage, deduplication and counting rules. This makes it possible to audit every table and figure without redistributing licensed database exports.

What the analysis reveals

Three patterns made visible by the data

A more specific clinical frontier

Research has diversified beyond surface-temperature thermography. Within the human-health core, the cluster combining diagnostic imaging, breast cancer and artificial intelligence rose from 9.3% of thematic assignments in 1980–1999 to 25.7% in 2020–2025. Fever, infection screening and thermometry also gained prominence.

A broad yet centralised international network

The collaboration network included 130 countries and 1,293 connections. The United States, United Kingdom, Spain, France, Germany and China occupy particularly central positions. This is not a ranking of winners; it is a view of the international infrastructure through which knowledge circulates.

Scale, impact and inequality are not the same thing

The 20 most productive countries accounted for 78.2% of publications assigned to countries. Yet when citations are normalised by field and year, differences between major income groups become far less pronounced. Gini and Lorenz analyses revealed strong citation concentration, including within each income stratum.

Why is this Data Science & Analytics work?

Because the conclusions depend less on a single visualisation than on the quality of every decision that precedes it. It requires defining rules for merging databases, handling incomplete metadata, comparing indicators with compatible units, testing counting choices and distinguishing robust signals from small-sample effects.

This methodological discipline enabled the study to separate three dimensions that are often conflated: participation (who publishes and at what scale), relative impact (how a work performs against its field and year), and concentration (how citations are distributed). That distinction supports a fairer and more actionable scientific interpretation.

Rigor also means communicating limitations

A responsible analysis does not present its results as context-free truths. The article records, among other points:

  • some records did not contain an identifiable country and were therefore excluded from national and collaboration analyses;
  • normalised indicators use reference sets built within the study corpus itself;
  • the broad search strategy also captured neighbouring literatures, which were retained and identified rather than hidden;
  • very small themes are more sensitive to algorithmic choices, so conclusions focus on macrostructure and dominant clusters.

Science in every line

Well-structured data expand the reach of science

By combining bibliometrics, network analysis and reproducible curation, the collaboration with Takotsubo Life Sciences turns an extensive literature into a rigorous map—useful to researchers, healthcare services and institutions deciding where to focus their next questions.

Read the published article · Access the reproducibility package