<p>In the era of big data, the development of efficient methods for predicting and discovering novel luminescent materials is of great importance. This work presents a data-driven approach to correlate the photoluminescence properties of activated ions with the crystal structure of the host lattice in doped inorganic compounds. Using RbBaPO₄ doped with Sm³⁺, Eu³⁺, and Eu²⁺ as an example system exhibiting excellent luminescence, we demonstrate a comprehensive fitting methodology to decode spectral features and site occupancy. The samples were synthesized via a conventional high-temperature solid-state reaction. The lattice occupancy of dopant ions was analyzed in detail using preferential occupancy site theory (POST). Photoluminescence excitation and emission spectra were accurately fitted using Gaussian functions, yielding key parameters that showed high consistency with experimental measurements. This study provides a generalizable framework and essential data support for future high-throughput prediction and design of luminescent materials based on spectral fitting and structural analysis.</p> Graphical Abstract <p></p>