Deucalion: A dataset for flood related research.

Flood disasters have been occurring with increasing frequency in recent years, and climate change has been added as a contributing factor. It is estimated that the annual economic losses due to flooding reach 250 billion globally (Jogman 2021). Disaster management aims to mitigate the negative impact of these events. It is a broad scientific, research, and operational field involving many different disciplines. Over the last two decades, new technologies and, more recently, artificial intelligence have emerged with innovative solutions that aspire to contribute effectively. This publication introduces a flood dataset consisting of specific flood-related captures from Instagram, primarily regarding the Ianos medicane (mainly in Greece, September 2020) and from two Kaggle sources. This mixed set of photos was imported into Label Studio, and various features were extracted. The exact methodology used, along with the dataset details, can be found in this manuscript.