Cloud Incident Data Analytics: Change-point Analysis and Text Visualization

When security incidents occur in a cloud computing environment, it constitutes a wake-up call to acknowledge potential threats and risks. Compared to other types of incidents (e.g., extreme climate events, terror attacks and natural disasters), incidents pertaining to the cloud security issues seem to receive little attention from academia. This study aims to provide a starting point for further studies via analytics. Bayesian change-point analysis, often employed to detect abrupt regime shifts in a variety of events, was performed to identify the salient changes in the cloud incident count data retrieved from Cloutage.org database. Additionally, to get to the root of such incidents, this study utilized text mining techniques with word clouds to visualize non-obvious patterns in the summaries of cloud incidents. Both quantitative and qualitative analyses for exploring cloud incident data offer new insights in finding commonality and differences among the causes of cloud vulnerabilities over time.