Correlation clustering and manifold learning, traditionally seen as distinct tasks, possess underlying similarities often overlooked. This paper elucidates on some of these shared traits, as well as on differences, focusing on local principal component analysis (LPCA) as a correlation clustering method and locally linear embedding (LLE) as a manifold learning approach. We elaborate on how both methods capture intrinsic data structures and handle complex distributions, suggesting convergence in objectives. While the findings are not generalizable over all manifold and correlation clustering methods, the gained insights serve as a basis for future investigations.