Projects

Centralized collective decision-making in nest-hunting ants hinders accuracy but facilitates consensus by decreasing effect group size

Marco Fele, Sara Ghivarello, Eleonora Gatti, Eva Rifà Rovira, Giulio Bettio, Lorenzo Zaffina, Carlotta Nunzi, Francisco Javier Marín Rodríguez, Bianca Pacini, Enrique Rozas Garcia

Two extremes of collective decision-making exist. In decentralized decision-making, all individuals contribute equally to the collective choice, analogous to “democratic” decisions in human contexts. In contrast, centralized decision-making occurs when some individuals contribute disproportionately, resembling “oligarchic” decisions where a few leaders exert disproportionate influence. Surprisingly, the consequences of centralized decision-making, and whether it may confer functional advantages, remain poorly understood. In Temnothorax ants, previous studies have shown that nest-site selection is oligarchic: a small subset of leaders disproportionately drives information transmission. In this project, we investigate the general consequences of centralized decision-making and link our theoretical findings to ant behavior to demonstrate relevance in natural collective-choice scenarios. We hypothesize that leadership reduces the “effective group size,” thereby decreasing decision accuracy through the wisdom-of-crowds effect while increasing consensus via noise-induced order. We develop a rigorous metric for quantifying effective group size and show that this metric predicts accuracy and consensus across models of varying complexity. Finally, parametrizing our model with empirical ant data, we demonstrate that colonies might trade-off accuracy to achieve higher consensus.

How voting in national and European Parliament level among parties influence each other

Sourin Chatterjee, Riccardo Sbarbati, Beatriz Arregui, François Gu, Mojtaba Roshana, Francesco Zambelli

How do voting behaviors at the national and European parliamentary levels influence each other across political parties? We study the entanglement of voting patterns between national parliaments and the European Parliament by analyzing votes in the European Parliament and in six European countries: Sweden, the Netherlands, Italy, France, Germany, and Poland. Using a quantitative disagreement metric between political parties, we identify distinct patterns of inter-party alignment across countries and between the two parliamentary levels. In parallel, we measure intra-party cohesion by analyzing voting similarity among Members of Parliament (MPs) within the same party, again observing substantial cross-country variation. Our results highlight significant diversity in parliamentary behavior at both national and European levels, suggesting a diverse range of interactions between these two levels yet to be understood.
These findings open future research on studying how voting relationships shift across topics, contexts, events, and time, and on linking the multilayer network to models of opinion dynamics.

Career trajectories in electronic music

Clara Eminente, Elena Candellone, Eider Pérez-Ordoyo, Juan Antonio García Castillo, Francesco Bertolotti, Martín F. Díaz, Duncan Cassells

This project's main aim is to predict the future success of electronic music artists from the events that they have played in order to identify behaviour patterns that may explain career trajectories and uncover profile-building venues. To measure individual success, we gathered Spotify popularity scores, Resident Advisor follower counts , as well as their Wikipedia page popularity as common metrics used in different fields. Success metrics are later combined with a scraped dataset of events from Resident Advisor modelled as a bipartite network of artists and the venues at which they played, allowing measurement of centrality and first steps towards predictive regression and gradient descent methods.The project hence uses three new datasets describing careers in electronic music, and presents an initial exploration of methods and features that may predict success based on the number of events they attended to.

Attenuating contagion in the global trade network

Joe Simpson, Emile Emery, Malvina Bozhidarova, Maria Sahakyan, Gabriela Retamales, Francisco Malveiro, William Funez Izaguirre

We study how adaptive trade behavior shapes the propagation of shocks in the global trade network. Specifically, we consider scenarios in which countries actively modify their trade relationships once a trading partner becomes affected by an adverse shock. Such adjustments, often interpreted as risk-mitigation or resilience-enhancing strategies, can substantially reconfigure the network of international trade. Using a set of different strategies, we analyze how endogenous rewiring of trade links influences both the structure of the network and the dynamics of shock transmission.

Connected by Science. Exploring Scientific Distance in the WWCS Network

Maria Sahakyan, Nelson Aloysio Reis de Almeida Passos

How closely connected are the scientists participating in WWCS 2026? Prior research on social and communication networks has consistently found that the average shortest path between individuals is approximately six hops — a pattern first observed in Milgram’s “six degrees of separation” experiment and later formalized by the Watts–Strogatz small-world model. Using publicly available OpenAlex data, we constructed a co-authorship network including workshop participants with indexed publications and their first- and second-degree collaborators (V = 80,697; E = 104,304). The resulting network is substantially more compact than the commonly observed six-hop pattern, with an average path length of 4.5 hops and a maximum distance of seven hops between any two participants. Overall, these findings indicate strong structural cohesion within the WWCS 2026 research community and suggest favorable conditions for rapid knowledge diffusion and future collaboration.

Searching for hidden paths towards zoonotic risks

Gabriella Cruz, Juliane Teixeira de Moraes, Lucía Cano de Arriba, Adrián Gutiérrez Arroyo, Kevin Teo, Marta Grasa, Laia Barjuan Ballabriga, Enahu Tahitu

Zoonoses are diseases caused by parasites transmitted between animals and humans, posing a growing public health concern. Many zoonotic parasites are transmitted by arthropod vectors, such as mosquitoes and ticks, giving rise to vector-borne diseases. The distribution and abundance of arthropod vectors are strongly influenced by land-use change. Among the various processes that drive these changes, including urbanisation, agricultural expansion, and deforestation, the role of wildfires and associated air pollution, particularly PM_2.5 concentration, remains underexplored in the context of zoonotic diseases. Fine particulate matter (PM_2.5) is a key indicator of air pollution linked to biomass burning and can also reflect broader ecological disturbances. Exposure to PM_2.5 is associated with physiological stress in human populations and may indirectly increase susceptibility to infectious diseases. Ecologically, wildfires can disrupt habitats, alter host and vector distributions, and modify human-wildlife contact rates. Unlike most land-use changes, which lead to long-term structural transformations of landscapes, wildfires represent short-term but potentially severe disturbances. Although currently smaller in magnitude than long-term land-use effects, climate change is expected to increase wildfire frequency and intensity. The central question of this study is whether wildfires affect the incidence of vector-borne diseases. We hypothesize that wildfires, by temporarily removing wildlife shelter and altering ecological conditions, increase human exposure and the spread of zoonoses. We analyzed data from Spain and the Brazilian Amazon. For Spain, we compiled information on agricultural and forestry land cover, livestock numbers, fire incidence, PM_2.5 concentrations, mammal and vector species richness, and vector-borne disease incidence in humans across 17 autonomous communities over three years. For the Brazilian Amazon, we examined vegetation cover, fire incidence, PM_2.5 concentration, and human disease incidence across 7,424 municipalities over ten years. In Spain, preliminary results show a negative effect of forestry cover on fire incidence and a negative association between fire, PM_2.5 concentration, and vector-borne disease incidence. In contrast, in the Brazilian Amazon, PM_2.5 and forestry cover were positively associated, and both have increased vector-borne disease incidence. These contrasting patterns could reflect differences in forest structure and land-use dynamics. Although the Amazon biome is by no means homogeneous, its forests are generally denser than those found in Spain, where the forest cover is predominantly low-canopy dehesas. The biome also supports a wide range of human activities, including anthropogenic wildfire, aggressive agribusiness expansion, and deforestation, increasing human exposure to zoonotic parasites. Overall, higher temporal and spatial resolution data would improve our ability to capture the effects of fire on land-cover change, thereby strengthening the evidence base for land-protection policies relevant to human health. As a next step, exploring how wildfire effects may evolve under future climate scenarios would inform policies aimed at protecting natural habitats and reducing the risk of increasing zoonotic spread.