Software defect datasets play an important role in evaluating software engineering techniques, including tools for fault detection, test generation, and automated program repair. The goal of this tutorial is to discuss the state of the art in software defect datasets, and to provide an opportunity to learn about BugSwarm, a large-scale dataset with thousands of reproducible software defect artifacts. The tutorial will consist of a 60-minute lecture and a 120-minute hands-on session. The lecture will give an overview of the state of the art in software defect datasets, discuss important topics related to the use of software defect datasets, and describe the BugSwarm dataset. The hands-on session will walk the participants through the process of selecting artifacts from the BugSwarm dataset based on specific evaluation tasks, running popular tools on the selected artifacts, and interpreting their results.