We introduce EventNarrative, a knowledge graph-to-text dataset from publicly\navailable open-world knowledge graphs. Given the recent advances in\nevent-driven Information Extraction (IE), and that prior research on\ngraph-to-text only focused on entity-driven KGs, this paper focuses on\nevent-centric data. However, our data generation system can still be adapted to\nother other types of KG data. Existing large-scale datasets in the\ngraph-to-text area are non-parallel, meaning there is a large disconnect\nbetween the KGs and text. The datasets that have a paired KG and text, are\nsmall scale and manually generated or generated without a rich ontology, making\nthe corresponding graphs sparse. Furthermore, these datasets contain many\nunlinked entities between their KG and text pairs. EventNarrative consists of\napproximately 230,000 graphs and their corresponding natural language text, 6\ntimes larger than the current largest parallel dataset. It makes use of a rich\nontology, all of the KGs entities are linked to the text, and our manual\nannotations confirm a high data quality. Our aim is two-fold: help break new\nground in event-centric research where data is lacking, and to give researchers\na well-defined, large-scale dataset in order to better evaluate existing and\nfuture knowledge graph-to-text models. We also evaluate two types of baseline\non EventNarrative: a graph-to-text specific model and two state-of-the-art\nlanguage models, which previous work has shown to be adaptable to the knowledge\ngraph-to-text domain.\n