•Novel empirical hysteresis model based on a data-model driven approach.•First data-driven approach in modelling the hysteresis of timber-based structures.•The experimental cyclic response is needed for the model formulation.•A step-by-step optimization problem defines the evolution of the backbone curve.•The maximum displacement during the cycle is the hysteretic evolutionary parameter.•Publication of the MATLAB and Python code implementations of the hybrid hysteresis model. This paper presents a novel computational approach to empirical hysteresis modelling applied to timber-based structures based on a combined data model-driven strategy. While the backbone curve is simulated using the experimental cyclic response based on a step-by-step optimization problem (data-driven approach), analytical functions describe the re-loading curves (model-driven approach). Empirical hysteresis models developed so far for timber structures are model-driven. However, the backbone curves can exhibit a highly irregular non-smooth trend, difficult to mirror using analytical formulations. The challenge in mirroring the experimental backbone using closed-form formulations has led to an extended set of parameters to be calibrated in existing literature models This paper presents a novel approach to the empirical hysteresis model, where the experimental data are directly involved, as a whole, in the model formulation. This model aims to be a possible trade-off between model complexity and accuracy. A reduced number of parameters needed to describe the re-loading paths is counterbalanced using an entire subset of the experimental data. The paper delivers the developed Matlab and Python codes for further implementation as a user-defined element within a Finite Element software.