In the world of citations, are we really OPEN(minded), or stuck in a comfort zone?
Mapping the Flow of Italian Science is an interactive platform that visualizes the results of a research project on Open Science.
The project builds on a recent study that created comprehensive datasets from the publication records of Italian universities managed through IRIS, the research information system adopted by Italian academic institutions. The datasets include all IRIS publications from selected universities that are indexed in OpenCitations Meta, together with all citation relationships recorded in OpenCitations Index involving those publications, either as citing or cited works.
Building on this foundation, we collected and analyzed additional data to investigate citation networks and knowledge flows across six Italian universities: UNIBO, UNIMI, UNITO, UNIPD, SNS, and UPO.
- RQ1: Which institutions and countries cite the publications of these universities, and which one are cited by them?
- RQ2: Which disciplines cite the publication of these six Italian universities, and which disciplines do those universities cite in return?
Through a series of interactive visualizations, this website reveals patterns of research connections and disciplinary exchange, offering new perspectives on how knowledge circulates within and beyond Italian academia.
Explore the data and discover the flow of Italian science.
Research Question I
The Map of Italian Science
Every scientific publication is a node in a global network — it cites and is cited, connecting researchers across institutions, disciplines, and borders. But which borders? And do all Italian universities connect to the world in the same way, or does each institution carry its own geographic signature?
The Map of Italian Science traces the citation flows of six major Italian research institutions: the University of Bologna, the University of Milan, the University of Padua, the University of Turin, the University of Eastern Piedmont, and the Scuola Normale Superiore. For each institution, we mapped both directions: which countries and organizations cite their research (incoming), and which countries and organizations they cite in return (outgoing).
The result is a portrait of Italian science as a networked system — who it talks to, who talks back, and where the conversation is uneven.
The map shows the top 40 citation partners for each institution, in both directions. Use the controls to switch between institutions and between incoming and outgoing flows.
The first thing the data tells you is how similar the six institutions are. Regardless of size, location, or disciplinary focus, all six Italian universities draw from and contribute to the same core group of countries.
This is not a coincidence — it reflects the structure of global science itself, where a small number of countries concentrate the majority of research output, journal infrastructure, and citation activity.
Six countries appear in the top 12 citation partners of every institution, in both directions: the United States, France, the United Kingdom, Germany, China, and Spain.
Each bubble is a country.
Size reflects
citation volume. Gold bubbles appear in
every
institution's top 12 — click to see which institutions share them.
Below the top six, the shared skeleton begins to fracture.
Japan appears for the four
large universities
— but UPO replaces it with India, SNS
with
Russia. South
Korea and Turkey appear only for the
two smallest institutions.
Click on a country to see which institutions share it as a top citation partner.
The shared skeleton is real — but knowing which countries appear tells you nothing about whether the exchange is equal. It isn't.
The relationship with the United States is unbalanced. Across all six institutions, Italian universities cite American science far more than American science cites them back.
China inverts the pattern. Chinese publications cite Italian research more than Italian research cites Chinese publications back — likely reflecting Western-centric citation practices and disciplinary differences between the two scientific communities.
France is the exception: a genuinely balanced partner. Citation flows in both directions are roughly equal across all six institutions — a sign of real bilateral collaboration, rooted in shared European research infrastructure.
Each institution cites the world differently.
This chart maps those differences — not in absolute terms, but
relative to the
average across all six Italian universities.
Italy exports knowledge to the world
For most countries — especially across the Global South — the map turns yellow. They cite Italian research more than Italy cites them back.
Italy cites America. America doesn't always cite back.
Across all six institutions, the US shows an outgoing bias — except one.
Scuola Normale Superiore is always the outlier
The only institution nearly balanced with the US. Over-indexed toward Russia, under-indexed toward China.
University of Eastern Piedmont looks East
The most Asia-oriented institution in the group — consistently, across both charts.
Moving from countries to organizations reveals a similar structural
convergence: all six Italian institutions engage with the same core set of international
partners.
A deeper analysis of the data uncovers additional patterns.
The visualizations below show the top 15 organizations that either cite, or are cited by, the six Italian institutions included in this study. They also reveal the degree of reciprocity in the citation relationships.
Italian organizations occupy many of the top positions. To better understand each institution's international citation profile, the analysis can be viewed both with and without Italian organizations. When Italy is included, domestic citation networks appear dense and highly reciprocal. The near-perfect symmetry of these relationships suggests that citations among Italian institutions are mutual rather than driven by one-sided dependence.
Removing Italian organizations from the analysis highlights important differences between institutions. UPO undergoes the most striking transformation: its previously concentrated network gives way to a profile centered on large-scale physics infrastructures and research organizations, for instance including CERN and IN2P3. In this configuration, its citation pattern becomes structurally similar to that of SNS.
Two organizations stand out across the dataset: CNRS and Harvard University Press. CNRS records the highest overall citation volume among all organizations connected to the six Italian institutions. Its position almost exactly on the diagonal line indicates a highly balanced relationship: Italian universities cite CNRS as much as CNRS cites them. This combination of prominence and reciprocity is remarkably consistent across institutions.
United States organizations are consistently positioned above the diagonal, indicating that Italian institutions draw more heavily on American research than American organizations cite Italian work in return. Harvard University Press represents the clearest example of this asymmetry, making publisher-driven citation dependence one of the most visible patterns in the data.
By contrast, Chinese organizations tend to appear below the diagonal. Across all six institutions, they cite Italian research more frequently than Italian institutions cite them.
Taken together, these patterns reveal distinct modes of scientific exchange. European organizations generally appear as reciprocal peers, United States organizations as dominant reference points, and Chinese organizations as increasingly active - but still asymmetrical - partners in the global circulation of knowledge.
A country’s importance is not only a matter of volume. It also depends on how that volume is distributed. Some countries spread their citation activity across dozens of organizations, creating dense and resilient networks. Others rely heavily on one or two institutions, making their connections more concentrated - and potentially more fragile.
The sunburst charts reveal this difference at a glance. Wide, balanced rings indicate a diversified ecosystem; large dominant slices point to institutional dependence.
Large scientific systems— such as the United States, France, and Italy —tend to distribute citations across dozens of organizations. With the notable exception of CNRS, no single institution dominates. These relationships are systemic rather than bilateral, making them more resilient: the removal of any one organization would have little effect on the overall pattern.
Smaller countries with lower citation volumes, such as Switzerland and Belgium, often display the opposite structure. A large share of their citation activity is concentrated in one or two organizations. These connections can be highly valuable, but they are also more dependent on specific institutions than on the broader national research system.
This distinction matters for research strategy. Citation volume alone does not necessarily indicate deep integration. A country generating thousands of citations through a single organization may be structurally more vulnerable than a smaller partner whose citations are distributed across a diverse network of institutions.
Research Question II
Disciplinary Flows
Which scholarly disciplines cite the publications of these six Italian universities, and which do they cite in return? To answer this, we traced over 80 million citation records across all institutions and mapped them onto a shared classification system.
LOC Main Classifications *
Disciplines are based on the Library of Congress (LOC) Main Classifications, a broad internationally recognized scheme.
Additional Categories — not part of the LOC Main Classifications; added during preprocessing
* LOC Main Classifications: Library of Congress Classification Outline
Now let's look at how these disciplines appear in the data.
What disciplines appear most in citation activity across the six institutions?
Science and Medicine account for over 60% of all citation activity, but every institution tells a different story.
Across all six institutions, two disciplines dominate: Science (≈32%) and Medicine (≈31%). The remaining categories — Geography/Anthropology, Technology, Agriculture, and others each contribute between 2–7%. Yet beneath this shared dominance, the balance shifts from one university to the next.
- UNIMI : Medicine-heavy. Medicine leads both citing and cited, with the largest gap between the top two disciplines and the rest of any institution.
- SNS : Science-dominant. Science alone accounts for ~50% of citing and ~56% of cited activity. Medicine is a distant second.
- UPO : Science and Medicine equally dominant, making UPO the most concentrated Science–Medicine institution overall.
- UNITO : Shifts toward Medicine on the cited side, with Agriculture as a notable secondary discipline.
- UNIBO & UNIPD : Science-led but the most broadly distributed, with visible contributions from Technology, Geography, and Social Sciences.
Discipline share of citation records
Citing / Cited volume by discipline
Hover over any bar to see exact values. Citing = disciplines that cite this institution's publications. Cited = disciplines cited by this institution's publications.
These counts show which disciplines appear. Next, we trace the flows: who cites whom.
When disciplines cite, where do their citations go?
Large disciplines are self-contained.
Smaller ones tend to be evenly distributed while orbiting
Science and Medicine.
Each segment shows the share of a discipline that is citing the selected institution. Click any segment to drill down, and it will become the inner ring. The outer ring reveals what disciplines are the works they cite. Click the center to reset.
Click an inner segment to explore a discipline's citation flows.
Inner ring = citing discipline share. Outer ring = cited discipline breakdown. All flow types (Incoming, Internal, Outgoing) combined.
The sunburst shows the aggregate picture. Now, trace the flows institution by institution.
How do citation flows differ across institutions and flow types?
Trace the exact path of citations —
from discipline to discipline.
Left nodes = citing disciplines. Right nodes = cited disciplines. Link width = citation volume. Hover for exact counts.
Every filter is a new question. The data is yours to explore.
What the data tells us
Citation networks are not neutral maps of knowledge — they reflect the organisational structures, disciplinary hierarchies, and geopolitical asymmetries of the research system itself. Across all six institutions, citation activity concentrates around a small number of countries and disciplines.
Science and Medicine dominate everywhere. The United States, France, the United Kingdom, Germany, China, and Spain form a shared core that no Italian university escapes.
But the data also reveals something less expected: for most of the world, Italian science is cited more than it cites back. The asymmetry runs deeper than a simple centre-periphery story. It is shaped by language, disciplinary tradition, publishing infrastructure, and institutional size — none of which citation counts alone can explain.
What we don't know yet
Citation patterns can show us where knowledge flows. They cannot tell us why it flows the way it does.
Why are we always looking at the same countries? Why do some barely appear at all? What if some of the most important connections are waiting to be made — not within a single discipline or country, but across them?
Citation is far more than a technical academic practice. It is the channel through which scientific knowledge is absorbed, evaluated, and extended. But it can also reinforce boundaries — between disciplines, countries, and communities. Researchers tend to cite within familiar networks, creating stable but limited ecosystems of knowledge exchange. Over time, these patterns risk reproducing the same perspectives rather than expanding them.
These questions point toward future research — into collaboration networks, funding structures, language barriers, and the role of open access in reshaping citation geographies.
Open science, open questions
The availability of open data made this analysis possible — but it also exposed
how uneven that openness can be.
This is why studying
citation patterns matters.
The circulation of knowledge does not automatically become more open just because
the infrastructure to study it does. The map is not
the territory.
But you can't change what
you can't see.
The Project
This project builds on the work of Andreose et al. (2026), who mapped
the coverage of
Italian publications in OpenCitations
and made the resulting dataset publicly available on Zenodo.
That dataset — linking IRIS publications from six Italian institutions to their citation
relationships in OpenCitations Index — formed the foundation
for everything you've explored here.
All code, data, and analysis pipelines are openly available in our GitHub
repository and on Zenodo.
If you want to go deeper into the methodology, findings, and implications, the full academic
paper is available here.