Our group is using the 3D library to graphically visualize co-occurrence networks within ALCIDE. The idea is to combine co-occurrences together with information about polarity of words involved in the network.
For Kennedy’s and Nixon’s speeches we have used SentiWords, a resource provided by FBK containing 155,000 words associated with a sentiment score.
Our main effort today (and not only today!) is on the ALCIDE (Analysis of Languange and Content In a Digital Environment) platform.
The goal of this project is to enhance historical research, in particular the analysis of political discourse, with state-of-the-art natural language processing technologies. In the early stage of the project, we focused on the writings and speeches of Alcide De Gasperi, one of the founders of the Italian Republic. More recently, we have moved to the automatic analysis of 1960 US presidential campaign, processing Kennedy’s and Nixon’s speeches.
A picture is worth a thousand words. Therefore, as an example of ALCIDE’s functionalities, we show below the plot of the different locations covered by the candidates during the presidential campaign. Can anyone explain why Nixon visited only few times the West Coast compared to Kennedy?
Based on the place of publication of De Gasperi’s public documents, ALCIDE can also show the places he visited more frequently during his activity. Enjoy!
In addition to the main task on contemporary news stories, we propose a pilot task on temporal processing of historical texts to promote the application of NLP techniques to the Digital Humanities domain.
For this reason Rachele is annotating De Gasperi’s writings about the outbreak of World War 1 with our tool CAT (Content Annotation Tool) following the It-TimeML specifications.
Annotated texts will be used to evaluate the performances of automatic Temporal Processing tools.