Textual data visualization for newsrooms

点击翻译按钮获取中文摘要

News media is going through a severe recession since the beginning of the 2000s. The newspapers sector has**seen its advertising based revenues dropped and many newsrooms had to shut down. Unlike the traditional**newspaper sector, the digital news content sees an increasing audience. Although the business model is still**under-development, this platform offers a huge potential in terms of audience and means of story telling. As in**any competitive industry, innovation becomes one of the important lever for each media to hold its own and**survive. Among the new practices presently being investigated, interactive data visualization is on the top list.**This practice is fostered through the increased potential and democratization of data science techniques.**Political discussions and electoral campaigns are among the most regular and popular topics that newsrooms**cover. These topics also produce a lot of open access textual data. However, this type of data offers challenging**issues in this context since they are often massive, hence hard to navigate and not easily and fully accessible to**both the news readers and journalists. Our objective through this collaboration with the Quebec media Le**Devoir, is to propose three innovative sets of tools that target three specific challenges dealing with different**types of textual data: parliamentary discussions, political debates, and twitter statements from campaigning**politicians. Each of these data vary in size, structure and visualization goal.