COVID-19 studies reveal negative correlations between computer-mediated communication and well-being, highlighting the role of face-to-face interaction in understanding human behavior. We present an IoT-based framework for studying face-to-face interactions in which participants use microcomputers to interact via Bluetooth signals that represent their social closeness. Participants adjust their interaction strategy in real time via a button on the device, which is displayed on the screen along with the accumulated score. We simulated a situation similar to the Social Particle Swarm model and its online (web-based) experimental variant, where formation and collapse of cooperative clusters emerged. This paper presents a framework and validates its performance through three experiments: one focused on capturing social relationship dynamics, the second comparing behavioral patterns between face-to-face and web-based conditions, and the final experiment testing the framework with a larger group. Implications for the importance of information visibility of others for active and cooperative social dynamics will be discussed.