Navigating Storms and Social Network research: ODISSEI Conference

As the storm clouds gathered ominously outside, I found myself standing in the midst of another kind of tempest—the whirlwind of ideas and insights at a recent conference where I had the honor of being a presenter. The day may have been fraught with weather-induced chaos, but within the confines of the conference room, a different kind of energy pulsed through the air—one of collaboration, discovery, and a shared passion for agent-based modeling (ABM). The realization that ABM could be applied to a myriad of future research topics and the ongoing development of its applications became a prevailing theme throughout the day.

In the vibrant discussions that unfolded, I discovered a rich tapestry of topics that could benefit from ABM applications in the future. The scope of ABM extends far beyond what I had initially perceived, and the conference became a melting pot of ideas, where professionals discussed the nuanced development and application of ABM in various domains.

The highlight of my conference experience, however, was the chance to present my own research on building an ABM to model the diffusion of vegetarian diets through social networks. In a world where the adoption of vegetarian diets holds immense benefits for both human health and the environment, understanding the mechanisms of diffusion through social networks is a crucial puzzle to solve.

In the context of my research, I underscored the scarcity of evidence regarding how vegetarian diets diffuse through social networks, contrasting the extensive studies on negative health behaviors. Leveraging ABM, I aimed to simulate the multifaceted factors influencing the diffusion process: media, values promotion, interpersonal influence, and the nuanced prediction of how vegetarian diets diffuse in diverse settings.

What sets our ABM apart is its enrichment with demographic microsimulation. Acknowledging the limitations of existing models that overlook essential demographic variables in measuring interpersonal influence, our approach strives for a more comprehensive representation of social networks.

The initial findings from my research underscored the crucial need for refining network generation models. It highlights the significance of constructing a social network that not only mirrors real-world dynamics but is also tailored to suit the specific phenomenon under investigation—in this case, the diffusion of diets. Attributes of the network, such as the maximum size of each personal network or the average personal network size, must be carefully considered to guarantee the authenticity of the simulated social structures. Failure to account for these factors could result in the swift dominance of particular groups, ultimately distorting the simulation results. However, the journey doesn't end with a presentation. The ongoing work in refining our model involves testing it against more reliable data sources and thorough relevant literature review. Calibration, a crucial aspect of our continued efforts, is focused on aligning our model with empirical data cohorts and existing literature, ensuring that our simulations remain grounded in reality.

As I concluded my presentation, the storm outside seemed to echo the turbulence within the world of ABM. Yet, amidst the atmospheric chaos, a sense of clarity emerged. The conference had not only provided a platform to share my research but had also opened doors to discussions, collaborations, and a deeper understanding of the evolving landscape of agent-based modeling.

ODISSEI became more than just a conference; it was a convergence of minds, a celebration of intellectual curiosity, and a testament to the resilience of shared ideas in the face of external storms. As the tempest eventually subsided, the echoes of insights and connections forged during ODISSEI continued to resonate, leaving an indelible mark on the landscape of ABM research and its dynamic future.

Astrid Anh Pham is a junior researcher at the chair group Health and Society. She is an enthusiastic data explorer — whether crafting a code to decipher inquiries or architecting data solutions for multifaceted projects.