The use of interaction analysis for creating agent-based models could lead to more empirical simulations. This study focuses on pedestrian-street interaction during negotiations at a signal-controlled crossing. It examines which data from interaction analysis could be used in the development of artificial pedestrian societies. This article sets out a framework for structuring the data deriving from interaction analysis and demonstrates the process of developing an agent-based model by translating the acquired data into the model. The data collected through interaction analysis serves as input for the agent-based model. The structuring performed on the interaction analysis data is used to define the outcome variables for agent-based modelling. The study concludes by proposing an initial framework that describes the use of interaction analysis in simulations.