An Online Structured Political Event Dataset based on CAMEO Ontology
Sayeed Salam, Patrick T. Brandt, Vito D’Orazio, Jennifer S. Holmes, Javiar Osorio, Latifur R. Khan · 2020
Political activities and interactions between different global entities are becoming growing field for data-intensive computing with a wide scope of research opportunities for both social science and computer science researchers. This research needs to be carried out at a local (limited to a particular region) and global scale, often divided in temporal manner. It is also useful to have the most recently updated dataset for relevant analysis. For these purposes, we need timestamped, geolocaated structured information about political interactions. Keeping this in mind, we develop a datatset that complies with Conflict and Mediation Event Observation (CAMEO) ontology inspired by the ”who-did-what-to- whom” format. We use a distributed framework for data collection and processing that works in real-time with Apache Kafka and SPARK in order to process a global collection of news data in different languages (i.e., Spanish, Arabic) and generate those structured event data in real-time. We also provide an API for easy access to the data. In this paper, we describe how the data is represented, collected, and processed, how we generate the most up-to-date dataset with dynamic ontology extension, and how to access the data and possible analytical problems that can be addressed by building a model on the dataset.