The study, conducted by Daniel Delatin Rodrigues and Marco Grasso of the University of Milan-Bicocca and Giuseppe Calignano of the University of Bologna, examines Enel’s power plants in Civitavecchia, a long-standing fossil energy hub near Rome. The authors introduce what they term the deactivation network approach, a framework that treats conflicts over fossil infrastructure as relational struggles between two antagonistic logics of action. On one side stand practices of continuity and legitimation, through which a fossil network works to keep infrastructure running. On the other stand practices of destabilisation and disruption, through which a heterogeneous coalition attempts to shut it down.

The empirical heart of the research is a natural experiment of sorts. Civitavecchia experienced two contested fossil-to-fossil conversion cycles with opposite outcomes. Between 2000 and 2010, Enel pursued an oil-to-coal conversion that succeeded despite local opposition. Between 2018 and 2023, the company proposed a coal-to-gas conversion that was halted, with the plants slated for closure in 2025, although that closure has since been postponed under pressure from the energy-security concerns raised by the Russia-Ukraine war and a broader European rollback on climate ambition. Comparing these two episodes allowed the researchers to ask what structurally distinguished failure from success.

Methodologically, the team took an intentionally asymmetrical approach. They mapped only the deactivation network, treating the fossil-support network as a structuring context that shapes political opportunity rather than reconstructing it as a second network. Using event-based social network analysis, they defined ties between actors by co-participation in deactivation events such as meetings, demonstrations, strikes, and joint communiqués. They then triangulated these structural patterns with eight interviews, documentary sources, archival materials, and digital ethnography of online spaces. Enel itself declined to be interviewed or otherwise collaborate with the research.

The structural results are initially puzzling. The deactivation network remained almost unchanged across the two periods: 220 ties in 2000-2010 and 216 ties in 2018-2023, with nearly identical average degree, very low density, and 95 percent of possible connections absent. The two networks are, in structural terms, essentially equivalent. Yet the outcomes diverged dramatically. The key difference lay not in how dense the network was but in who belonged to it and how intensely they collaborated.

In the first period, the network was dominated by activists, who accounted for more than half of all components and occupied the top positions in degree centrality. Institutions, political parties, scientific and health organizations, and trade associations played marginal roles. Crucially, trade unions were entirely absent. Interpreted through the framework, this configuration produced destabilisation without disruption: activists effectively denaturalized fossil fuel use, mobilized moral and scientific arguments, and challenged the legitimacy of the oil-to-coal conversion in public discourse. But these efforts remained disconnected from the institutional and organizational arenas where continuity practices were actually enacted. Destabilisation was never translated into disruption, and the conversion went ahead.

The second period tells a different story. While activists remained central, the network expanded and diversified with the entry of trade unions, with seven participating components, alongside political parties, institutions, trade associations, a technical group, and a media company. The density-by-groups analysis reveals a dramatic reconfiguration: ties within the activist group collapsed from 96 to 18, while ties between activists and unions rose by 20. Unions occupied a bridging position with high inter-group connectivity. The analysis of tie intensity, measured by the number of joint event participations, showed that groups established markedly stronger relationships in the second period, and that certain actors with few direct ties achieved high eigenvector centrality by connecting to well-connected parts of the network, exerting indirect influence.

The qualitative evidence explains why the union entry mattered so much. Unions have historically defended jobs against environmental closure, embodying the classic jobs-versus-environment dilemma. In Civitavecchia, however, the interviews and documents show that unions did not merely join an existing coalition. They rearticulated health, labour, and territorial futures as a single just-transition problem and co-produced, with technicians and ecological committees, a concrete industrial alternative to the gas conversion, notably offshore wind tied to local shipyard and port employment. This discursive and projectual work dissolved the very opposition on which the fossil network’s legitimation rested, converting relational proximity into disruptive capacity. As the authors put it, the network analysis locates the hinge; the qualitative analysis explains why it held.

The study also clarifies what network measures can and cannot do. Metrics such as density, centrality, and fragmentation describe relational configurations but do not register the meaning or strategic intent of practices traveling through ties, nor asymmetries of power. Tie frequency is not equivalent to influence. The deactivation network approach therefore treats social network analysis as one of two complementary operations: mapping how the network was configured and how it changed, while qualitative materials interpret what circulated through those ties and how relational proximity became political and administrative effect. It is this combination, the authors argue, that moves beyond a simple protest-event count or a stand-alone ethnography.

The broader implications reach into debates on supply-side climate politics and infrastructural power. Fossil fuel infrastructures, the authors emphasize, are never merely material objects; they exist through the relations they support and actively organize asymmetrical conditions of visibility, access, and intervention. They concentrate power while dispersing responsibility, normalizing continuity as a taken-for-granted background condition. Deactivation networks, by contrast, remain relationally exposed and politically contingent. The Civitavecchia case demonstrates that deactivation becomes plausible not when destabilisation intensifies in isolation but when it is relationally articulated with disruptive capacities, lawsuits, policy provisions, strikes, blockades, and alternative investments, capable of interfering with an infrastructure’s operational and institutional viability. For climate movements everywhere, the lesson is sobering and empowering at once: large demonstrations do not necessarily translate into substantial change, but broad alliances that reach into the arenas where permits and decisions are made can.

Subject of Research: Network dynamics of activist coalitions contesting fossil fuel infrastructure conversion in Civitavecchia, Italy

Article Title: From destabilisation to disruption: deactivation network dynamics in an Italian fossil infrastructure conflict

Article References: Delatin Rodrigues, D., Calignano, G., & Grasso, M. (2026). From destabilisation to disruption: deactivation network dynamics in an Italian fossil infrastructure conflict. Geographica Helvetica, 81(2), 347-360. https://doi.org/10.5194/gh-81-347-2026

Image Credits: AI Generated

DOI: 10.5194/gh-81-347-2026

Keywords: fossil fuel infrastructure, deactivation network, social network analysis, climate activism, energy transition, trade unions, just transition, Civitavecchia, Enel, coal-to-gas conversion, political opportunity, phase-out

The research, led by Xuyu Chen and colleagues at the Foshan Power Supply Bureau of Guangdong Power Grid Co., Ltd., introduces an active recommendation algorithm that combines two technologies: a domain knowledge graph and dynamic user personas. The goal is not simply to push more content at employees, but to anticipate which rights-and-interests services—benefits guidance, dispute resolution, safety programs, training opportunities—a given worker is likely to need, and to deliver those services before the employee has to ask. In their evaluation on real business data, the system achieved a Precision@10 of 83.5 percent and a Recall@10 of 81.6 percent, with an average response time of just 135 milliseconds.

To understand why this matters, it helps to consider what the existing approach gets wrong. Static rule matching treats an employee as a fixed profile: a job title, a department, a seniority level. But needs shift constantly. A field technician who recently logged a safety complaint has different immediate priorities from the same technician six months earlier, before an incident or a policy change. Manual processing, meanwhile, cannot scale across thousands of workers and dozens of service categories. The authors note that these limitations leave employees’ changing needs across job roles, policy contexts, and service scenarios largely uncaptured—a gap the new framework is designed to close.

The first pillar of the system is the knowledge graph. Rather than treating policy documents and service records as isolated text, the researchers construct a graph in which entities—employees, services, regulations, complaints, training programs—are nodes, and the relationships between them are explicit, machine-readable edges. Building this graph requires enhanced entity recognition, the natural language processing task of identifying meaningful terms in unstructured text, followed by relation extraction, which determines how those entities connect. Once assembled, the graph is compressed into knowledge graph embeddings: dense numerical vectors that encode the structure of the network so that a machine learning model can reason over it efficiently.

The embeddings are what make the graph computationally useful. In a raw graph, asking whether a particular grievance-handling service is relevant to a particular line worker requires traversing links and interpreting labels. In the embedded space, every entity becomes a point in a high-dimensional coordinate system where related entities sit close together. Similarity becomes arithmetic. A recommendation engine can measure the distance between an employee’s situation and a service offering, or trace which paths through the graph justify a match. That last property—traceability—is central to the authors’ emphasis on interpretability, because a recommendation that can point to the specific policy, complaint record, or role relationship that triggered it is far easier for union administrators to trust and audit than an opaque score.

The second pillar is dynamic user persona modeling. Where traditional user profiles are snapshots, personas in this framework are living summaries that capture recent changes in employee behavior. If a worker begins attending safety trainings, files a benefits inquiry, or interacts with complaint channels, those signals update the persona, shifting the vector that represents the person’s current needs. This recency weighting matters in an organizational setting, where a single event—a workplace incident, a policy revision, a seasonal workload change—can abruptly redefine what services are relevant. The persona acts as a bridge between the messy stream of behavioral data and the structured knowledge graph.

The recommendation engine itself is a hybrid. It fuses three signal sources: collaborative filtering, the classic technique that infers preferences from the behavior of similar users; the knowledge graph embeddings, which inject domain structure and policy context; and the persona features, which encode each employee’s evolving situation. Hybrid architectures of this kind are standard in modern recommender systems precisely because no single signal is sufficient. Collaborative filtering struggles with sparse data and cannot explain itself; knowledge graphs are rich in context but depend on the completeness of the graph; personas capture dynamics but can overfit to short-term noise. Combining them lets each component compensate for the weaknesses of the others, producing recommendations that are simultaneously personalized, contextually grounded, and explainable.

The evaluation used real business data from a power grid enterprise, spanning employee records, service transactions, complaints, safety events, training participation, and activity sign-ups. Before any modeling, the researchers standardized the data and applied privacy protections, fuzzifying sensitive fields and de-identifying personal information—a necessary step when the underlying records concern labor rights and grievances. Under a unified evaluation protocol, the hybrid model reached an F1-Score of 82.5 percent, balancing precision and recall, while maintaining a response time of 135 milliseconds, fast enough for interactive use. Those numbers suggest the system can operate not as a batch reporting tool but as a live service layer that responds to employees in real time.

The broader significance extends beyond one utility company. Organizational service recommendation is an underexplored cousin of the consumer systems that power streaming and shopping platforms, but it carries different stakes and different constraints. The data is more sensitive, the notion of relevance is tied to wellbeing and legal rights rather than engagement, and the users—union administrators and employees alike—need to understand why a recommendation appeared. By anchoring the system in an explicit knowledge graph and keeping the recommendation logic interpretable, the authors offer a template for how institutions can automate proactive support without sacrificing accountability. The open-access publication, released under a Creative Commons license, makes the full technical apparatus, including complete pseudocode and a summary of all formula symbols, available for other teams to examine and adapt.

There are, of course, caveats. The results come from a single enterprise’s data, and performance in other organizational contexts—with different service catalogs, record-keeping practices, and graph densities—remains to be demonstrated. The paper is also an early-access version, subject to further editorial revision before the final Version of Record. Still, the core finding stands: integrating structured domain knowledge with dynamic user personas measurably improves recommendation accuracy, coverage, and interpretability in organizational service settings. For the employees of one Chinese power grid, that could mean the right support arriving before it is requested. For the field of applied AI, it is another sign that recommendation technology is migrating from the marketplace into the workplace, where the currency being matched is not entertainment or products, but rights and care.

Subject of Research: A knowledge graph and dynamic user persona-based recommendation algorithm for proactive trade union services in power grid enterprises

Article Title: An active recommendation algorithm for trade union rights and interests services in power grid enterprises based on knowledge graphs and dynamic user personas

Article References: Chen, X., Song, C., Lin, H., & Li, K. (2026). An active recommendation algorithm for trade union rights and interests services in power grid enterprises based on knowledge graphs and dynamic user personas. Complex & Intelligent Systems. https://doi.org/10.1007/s40747-026-02486-y

Image Credits: AI Generated

DOI: 10.1007/s40747-026-02486-y

Keywords: knowledge graph, recommendation algorithm, dynamic user personas, trade unions, power grid enterprises, personalized services, collaborative filtering, knowledge graph embedding, machine learning, employee services, interpretability, Complex & Intelligent Systems