Linea: Tailoring timelines by visual exploration of temporal text

Timelines are often used to summarize complex, time-evolving, sequences of events. In this poster, we present the preliminary results of Linea, a tool designed to help users build timelines that summarize content from Wikipedia. We combine Natural Language Processing and the PageRank algorithm to automatically build default timelines of any Wikipedia article, giving users a starting point for tailoring them. Then, users can interact with the event matrix - a collection of histograms of events - by exploring each matrix cell in search for peaks, i.e., proxies of important periods of time. By selecting a peak, the tool shows all events in that period so users are able to add or remove events, thus customizing timelines.