Satirical News Corpus
Fake news is a form of news that consists of deliberate misinformation that is spread widely through digital media. This has been a growing issue in recent times especially with the advent of social media where people share posts without often verifying the legitimacy. Their sole purpose is often to mislead the audience and or cause widespread chaos. So, it has become increasingly important to identify fake news from legitimate ones. Although there is no aggregated collection of fake news, there are many websites around the world that post satirical news. We are hopeful that our corpus can act as an example of fake news. Satire and fake news have very similar properties, both use controversial misinformation to reach a wide audience. The corpus includes 11438 satirical news headlines from the ONION and 509 headlines were annotated using Mechanical Turk.
The code to the final corpus web app is stored here.
Please make sure you have the following python libraries installed before you try to use the app
flask
bs4
pandas
- Download or Clone the repo
- Go into the Production
- Run the following command
python deploy.py - Open the browser and visit the following URL -
http://0.0.0.0:8080/ - After you open the interface, please enter keyword(s) that you are interested in and click the 'Search' button on the top left. Some interesting keywords are 'Trump', 'Americans', etc.
- You can see the satirical part of the headlines by clicking the 'Show Annotations' text. (You have to select the text and not the checkbox. The checkbox will be automatically ticked when you click the text)