Chrissi's research proposal

[chendaniely]

@gitchrissi I fixed up your files and submitted the PR for you.

[datajonbrig]

I think the ideas are a great start, and are a strong reason to use the NewsAPI to pull huge numbers of news text to spot these trends. I'm going to paste each idea below and add some comments/feedback.

Ideas for research:
- Does the media intend to scare the public with their choices of diction in their articles

1. Intention will be hard to find/prove. You can use the language in articles to cluster around diction to identify which news source/author/political bias leaning creates the most positive/negatively charged diction around the virus. You could plot the sentiment or word choice change over time to see if certain sources were potentially fearmongering (i.e. source X was calling for martial law of Y country Z months before a case was even reported there, that seems emotionally charged). HOWEVER, truly being able to show that a person/source/site truly had a negative intention is difficult, and not something that can be done lightly.

2. Being scared. How will we identify which news articles are scary? Does that hold true for people of all ages, backgrounds, races, ethnicities, etc? If we can only identify scary to certain sub populations we'll need to note that

- Is there a difference between the scholarly articles vs the media articles reporting on coronavirus? What are the differences?

1. This is a very succinct question that has a clear goal and answer. You could perform cluster analysis based on word usage, and just on that alone we could see if scholarly/media are similar/different. Then we could add word change over time, sentiment, reach/impact, and further expand. Lovely question