Visual Representations of Temporal Relations between Events and Time Expressions in News Stories

High-quality annotation is essential for the effective predictions of machine learning models.When annotations are dense, achieving accurate human labeling can be challenging since the most used annotation tools present an overloaded visualization of labels.Thus, we present Vitra (Visualizer of temporal relation annotations), a tool designed for viewing annotations made in corpora, specifically focusing on the temporal relations between events and temporal expressions.This tool aims to fill a gap in the available resources for this purpose.Our focus is on narrative text, which is a rich source for these types of elements.Vitra was developed to increase the human capacity for detecting annotation errors and uncover relations between narrative components or issues about the annotation scheme.To show how this can be done, we present an analysis of a subset of the Text2Story Lusa corpus, a dataset of Portuguese news stories.Such analysis focuses on the linguistic properties of the events and temporal expressions that occur in the annotated texts, in particular, of short news.We highlight that annotation is an iterative process that involves multiple rounds of revision, and our tool facilitates this process by helping users detect inconsistencies and improve the annotation scheme, thus offering added value to the community.