Propose a project that uses a large, publicly accessible dataset. Explain your motivation for tackling this problem, discuss the data source(s) you are using, and explain the analysis you are performing. At a minimum, you will need to do enough exploratory data analysis to convince someone that the project is viable and generate two interesting non-trivial plots or other assets supporting this. Explain the plots and give url links to them.
Since the outbreak of the COVID-19 pandemic, numerous problems have arisen due to the spreading of fake news and misinformation. These problems are changing peoples behaviours and costing lives. I propose a look into how shared information about COVID-19 is shared and how this affects peoples' attitudes and behaviours, and consequently the spread of the virus.
I have conducted an investigatory analysis using time series data on the confirmed cases of coronavirus in the US, provided by the John Hopkins University. I have paired this with a Naive Bayes sentiment analysis of data from Official US Government Twitter account (state-level).
Twitter is a rich and up-to-date source of information that comes directly from the public. Twitter is also used by many official bodies to help share important information. It is therefore ideal for getting an impression of how information is shared and the effects this has on peoples' attitudes towards COVID-19.
- Confirmed Data: https://github.com/CSSEGISandData/COVID-19
- Death Data: https://github.com/CSSEGISandData/COVID-19
- Tweet Data: https://developer.twitter.com/en