Elezioni.io is an attempt to analyze and visualize the general Italian elections from a news standpoint. Each piece of news is analyzed through AI algorithms to understand the mood of an article (positive, negative or neutral) and the political areas, parties and people referred to. At the same time, we calculated the popularity of each article on Twitter and Facebook.
We have published several pages to tackle the topic from different angles. ‘Faccia a faccia con le elezioni’ (‘Face to face with the elections’) is a visualization of the latest news from the most important news outlets in Italy as perceived on the social networks. The reader can scroll through the latest 48 hours of news to have a first glance and later deep dive into the article.
We also published a more extensive view of the news with headlines and summary of the article with charts showing the evolution of popularity in the latest 48 hours.
In the ‘Analisi’ (‘Analysis’) section we have analyzed how the popularity of specific themes are related to the events happened along the election campaign. At the same time, we had fun trying to envision how the parliament would look like if the number of news and their popularity would have been the driving factor for the selection of the candidates.
Our plan is to keep producing contents on the past elections and on future ones. We will also apply the same methods to elections in other countries.
What makes this project innovative?
On the visualization side 'Faccia a faccia con le elezioni' takes a unique approach to show the news in the latest 48 hours, using 45 degree rotated squares as 'faces' with the mood of the article. The idea behind it is to give a clear idea of the general mood of the news on a specific day and the political area talked about.
What was the impact of your project? How did you measure it?
Source and methodology
All starts with a bot that is looking for tweets about the elections, among them, we collect the ones with links to articles in the most important news outlets. Each link is validated, explored and analyzed: its sentiment and topics are identified. After this step the link is pushed onto a queue that takes care of getting the change of popularity in time on Facebook and Twitter.
Bots and API are written in PHP with the help of Goutte, Twitter oAuth, Graph SDK and Wit PHP.
Frontend and datavisualization are based on React and D3.js