Visual Data Mining in Software Repositories: A Survey

Enormous quantities of data, collected and stored in large numerous data repositories, go unused or underused today, simply because people are unable to visualize the quantities and relationships involved. This huge amount of data has far exceeded our human ability for comprehension without powerful tools. Information visualization and visual data mining can help to deal with the flood of information. We can take advantage of visualization techniques to discover data relationships that are otherwise not easily observable by looking at the “raw data”. Visualization can add significant value when trying to understand not only the raw data available in large software archives, but also the results of data mining. These data are valuable especially in software maintenance activities, understanding software evolution and the socio-technical aspects of software development. Data mining and visualization are focal enablers for information recognition and knowledge discovery from any amount of data repositories. This paper present the results of a survey, which reviews some of the most common visual data mining (VDM) techniques and their usage in the software engineering field. The results indicate what kinds of aspects of the software engineering process are studied using VDM methods, and also the most common VDM methods used in the software engineering context.