Search Results

Knowledge graphs are the backbone of an increasing number of data-driven software solutions, where performance, i.e., runtime and accuracy, depends critically on the data they process. Consequently, synthetic knowledge graph generators become essential tools for reliable evaluation and future performance prediction of knowledge graph-driven solutions. These generators address critical challenges, including the exponential growth of knowledge graphs, the difficulty in acquiring real-world domain knowledge graphs, and the need for realistic data to predict future behaviour under realistic workloads. Generators overcome these challenges by producing scalable and controllable knowledge graph structures. Knowledge graphs at various scales enable the testing of existing solutions on larger-than-current graphs, supporting future-proof solutions. Controllable structures, in turn, support the generation of knowledge graphs in data-scarce domains and with realistic properties. This survey comprehensively analyzes the current landscape of synthetic knowledge graph generators, including evaluation tasks and metrics. We catalog 38 generators published between 2014 and 2025 based on their methodology, generation requirements, and the type of generated graph. We also survey the evaluation tasks and technical details reported for these generators, providing a reference for both practitioners looking for a generator and researchers designing new ones. We conclude with open challenges and promising research directions.