Scientists Discover Anti-Maskers Are Really Good At Science! Uh-oh.
Posted by:
John Rose
()
Date: May 20, 2021 06:01PM Viral Visualizations: How Coronavirus Skeptics Use Orthodox Data Practices to Promote Unorthodox Science Online
Scientists Discover Anti-Maskers Are Really Good At Science! Uh-oh. MIT did a study on Covid Skeptics and guess what? They found out we're better at science, data analytics and communication than they are! Link to study: [arxiv.org] Re: Scientists Discover Anti-Maskers Are Really Good At Science! Uh-oh.
Posted by:
Jennifer
()
Date: May 20, 2021 06:47PM Dear God - The Study is so complex. How much did that 'study' cost? Who paid for that?
[arxiv.org] One small part of 18-pages - 3.1.3 Network analysis. Finally, to analyze the users participating in these discussions, we constructed a network graph: nodes were users who appeared in our dataset, and edges were drawn between pairs of nodes if a user mentioned, replied to, retweeted, or quote-tweeted another user. The resultant graph consisted of almost 400,000 nodes and over 583,000 edges, with an average degree of 2.9. To produce a denser network structure, we calculated a histogram of node degrees and identified that two-thirds of the nodes were degree 1. We then computed subgraphs, filtering for nodes with a minimum degree of 2 through 10 and found that degree 5 offered us a good balance between focusing on the most influential actors in the network, without losing smaller yet salient communities. This step yielded a subgraph of over 28,000 nodes and 104,000 edges, with an average degree of 7.3. We detected communities on this network using the Louvain method [12]. Of the 2,573 different communities detected by this algorithm, we primarily focus on the top 10 largest communities which account for 72% of nodes, 80% of edges, and 30% of visualizations. Sorry, only registered users may post in this forum.
|
|