During an alarm flood, the alarm rate is greater than the operator can effectively manage. Many alarm data analysis methods have been proposed in the literature to mitigate the impact of alarm floods. This paper gives a review of the state of the art in alarm data analysis and aims at structuring the field. A distinction between sequence mining methods that apply to alarm sequences and time series analysis methods that apply to alarm series is suggested. The review highlights that online applications to help the operators during alarm flood episodes have been only treated as sequence mining problem in the literature to date. To address this gap, the paper also presents a binary series approach to classify ongoing alarm floods based on a set of historical alarm floods. The motivation for a binary series approach is demonstrated through an industrial case study of a gas-oil separation plant, and the performance of the presented method is compared with the performance of an established sequence alignment method.