<p>The purpose of data clustering in unsupervised machine learning is to identify meaningful patterns within datasets. Due to the extensive search space for patterns, deterministic approaches often do not work properly. Such problems can be considered NP-hard, as they typically involve nonlinear objective functions and unstructured search spaces. The goal is to find optimal centroids that minimize intra-cluster distances. Nature-inspired optimization algorithms are increasingly being adopted to tackle these challenges. Various algorithms are available to handle this type of problem. Among them, the Whale Optimization Algorithm (WOA) mimics the bubble-net feeding approach of humpback whales. This paper proposes a Lévy flight-based Whale Optimization Algorithm (LWOA) for data clustering. The algorithm is rigorously evaluated on fifteen well-known benchmark datasets and compared with eight well-known optimization algorithms. Performance is evaluated using the best, worst, mean, and standard deviation of fitness values. Based on the availability of ground truth in the datasets, the proposed algorithm is also evaluated and compared with existing algorithms with three metrics: normalized mutual information, adjusted rand index, and purity. The convergence behavior is evaluated using convergence curves, and the variability is analyzed using box plots. The effectiveness of the proposed algorithm is also verified using a statistical test. Three tests have been conducted: Friedman, Iman–Davenport, and Holm–Bonferroni tests at a 5% significance level. All experimental results demonstrate that LWOA consistently outperforms benchmarked algorithms in terms of cluster quality, convergence speed, and robustness. The code is available in the given link <a href="https://github.com/sahuashishcs/Hybridizing-Levy-Flight-and-Whale-Optimization-Algorithm-for-Effective-Data-Clustering">https://github.com/sahuashishcs/Hybridizing-Levy-Flight-and-Whale-Optimization-Algorithm-for-Effective-Data-Clustering</a>.</p>