Forced periodic operation can exploit inherent nonlinearities to exceed the steady-state performance. However, the application is limited by parameter selection and inadequate nonlinear analysis. To address these problems, the existing Nonlinear Frequency Response (NFR) method was proposed to judge the operation efficiency through the DC component. Nevertheless, the reliance on the first principle model and the incapacity of higher order nonlinearities analysis restrict the method to qualitative analysis. To address the limitations, a novel data-driven nonlinear frequency analysis method is proposed in this paper. It is to use the input-output process data and the selected numerical approximations to construct a data-driven NFR model, instead of the first principle model, to achieve quantitative analysis and the optimal design, which enables the maximal efficiency of the forced periodic operation. The method is then validated on an isothermal CSTR and a CO2 absorption device under square-wave forced periodic operation via both simulation and experiments respectively.