Powder diffraction experiments performed as a function of external variables like, e.g., temperature, pressure, or time open possibilities to gain additional information about the physical and chemical properties of the system under study. Sensible extraction of such information directly from in situ experiments requires the treatment of the dataset with methods like sequential and/or parametric refinements. We are progressing towards the development of the parametric refinement method, which performs simultaneous phase refinements of in situ powder patterns by imposing several rational physical models of the evolving parameters on the calculated powder diffraction profiles. One of the fundamental prerequisites for this method is that the powder patterns in the in situ dataset be grouped to their relevant phases. In this paper, we present an analytical method which uses the Pearson’s correlations coefficients of the powder patterns to automatically determine the phase transition points of the in situ powder dataset. The phase transition points determined are used to group the powder patterns belonging to identical phases and to prepare the patterns for automated sequential phase refinements. The proposed algorithm is implemented as an automated module in the multi powder diffraction pattern, data reduction software Powder 3D.