The EuropaBON Stakeholder Dashboard: A dynamic web application to map Europe’s biodiversity community

PLOS One | https://doi.org/10.1371/journal.pone.0329390 August 13, 2025 5 / 16 EuropaBON website and social media channels (X, https://x.com/EuropaBon_H2020, and Facebook, https://www.facebook.com/ EuropaBONH2020). While scheduling conflicts were not accounted for in the analysis, we recognize that availability constraints may have influenced participation. Additionally, funding was provided to support the involvement of certain members, thereby influ- encing the level of engagement of those members. Activity levels were assigned as absolute values (0–17) and categorized into three groups: 0 = none, 1 = low (participation in up to 8 events), and 2 = high (participation in more than 8 events). On the dashboard, different colors indicate activity levels: the darker the fill color of each symbol, the more activities the node participated in. Each node was assigned a position in the data flow (i.e., biodiversity data user, biodiversity data provider, or both) based on the information provided by each member during the registration process (S1 File). On the dashboard, the visualisation uses distinct shapes to represent different member institutions: circles for data users, triangles for data pro- viders, and diamonds for those who serve as both. Additionally, a separate group, “N.A.” (Not Available), includes nodes that are not registered members but were identified as data users or providers by members. In the network graph, these non-member nodes are displayed as grey dots. Web application architecture

The dashboard has been developed as a fully responsive JavaScript [15] web application using technologies that are licensed as free software using jQuery [16], the Bootstrap 4 framework [17] and various JavaScript libraries, such as Leaflet JS [18], Highcharts [19] and datatables.js [20], a plug-in for the JavaScript library jQuery. The network graph and centrality calculations were generated using Cytoscape.js [14], an open-source JavaScript library for graph visualisation, employing a force-directed FCOSE [21] layout to identify clusters and bridges (S6 File). The API interface between the web application and the server was developed in PHP 8 [22], which runs on the members portal’s [13] existing PHP/Mari- aDB stack using MariaDB v10.6.5 [23] as the database system. The entire website, including the database, is hosted in a VM Docker environment at iDiv (Germany). The deployed website uses Apache web server technology. The entire source code of the dashboard is available in the EuropaBON GitHub repository (https://github.com/Europa- BON/stakeholder-dashboard), and an archived version has been made publicly available on Zenodo [24]. Further technical specifications are available in the S7 File. For an overview of the web application architecture, see Fig 1. One limitation of the JavaScript library Cytoscape.js was its restricted functionality for advanced statistical analyses. To explore node relationships in more detail, we conducted additional network data analyses outside Cytoscape, which will be presented in the next section. Statistical model of network data

This section describes the statistical methods used to examine the factors associated with degree centrality and the number of activities in which EuropaBON members participated. Since both outcomes are count variables (i.e., non-negative integers), we considered a range of count regression models [26,27]. After evaluating several alternatives, we selected the negative binomial regression model as the most appropriate. This model accounts for overdispersion – where the variance exceeds the mean – a common feature in count data that can compromise the reliability of standard Poisson regression by underestimating standard errors and overstating significance levels. The negative binomial model mitigates this issue by incorporating an additional disper- sion parameter, offering a more flexible and robust alternative to the standard Poisson model [28]. In the estimated model, the variation in the number of events a member institution participated in, or alternatively, the number of connections between mem- bers and non-member participants in EuropaBON (degree centrality), was explained by a set of member-specific characteristics. Using the information provided by individuals in the registration process (S1 File), the following variables were included: position in the data flow, occupational sector, geographical region (distinguishing between countries within Europe that are members of the OECD – the Organisation for Economic Co-operation and Development – and non-OECD countries outside of Europe), realm, and

EU directives. The directives variable was based on responses to the question, “Which of these EU directives is of most interest to your work?”, which was included in the registration process to better understand the policy contexts informing participants’ exper- tise and focus areas. Although the responses reflect individual-level perspectives rather than institutional mandates, we included