Citation
Optimizing Social Media Sharing of Information Articles

Material Information

Title:
Optimizing Social Media Sharing of Information Articles
Creator:
Powell, Collin
Physical Description:
D2P Presentation

Notes

Abstract:
When we have such a large global event such as the outbreak of Covid-19, there is often a great deal of misinformation spread. This misinformation, conjecture, or unsupported commentary can lead to many negative impacts including economical impacts such as hording or social panic which can create massive mental health issues and a variety of other impacts. There have been several ways this has been combated with the current crisis, including search engines like Google prioritizing CDC or other government articles on their search results. However, Social Media is one of the major venues through which information is shared and viewed. It is much harder to implement technological changes to affect what articles individuals share with others. To help, an analysis of what aspects of online articles most impact how articles are shared is performed in order to provide policy guidelines for agencies to optimize article penetration.
Acquisition:
Collected for Auraria Institutional Repository by the Self-Submittal tool. Submitted by Collin Powell.

Record Information

Source Institution:
Auraria Institutional Repository
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.

Auraria Membership

Aggregations:
Auraria Library
Added automatically

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Optimizing Social Media Sharing of Information Articles By Collin Powell MATH 7393 with Joshua French

PAGE 2

Introduction and Motivation Improve Social Media penetration for Information releases on important topics. Major events, such as COVID, present a great deal of misinformation, misleading articles, and panic inducing commentary. Filters for such things can be limited or ineffective when shared between friends directly. Goal: Provide policy guidelines for official articles to improve share rates.

PAGE 3

Process and Data Data from the online archive Mashable ( www.mashable.com ) on article shares. Data is from 2015, well before the current pandemic crisis. Original set contains 61 tracked attributes, data cleaned and reduced to 20 for consideration. Source: http://archive.ics.uci.edu/ml/datasets/Online+News+Popularity Bayesian Linear Regression Analysis. Statistical analysis of the impact of each attributes effect on article shares.

PAGE 4

Attributes Considered Variable Description num_words_title num_words_article Number of words in the article average_word_length Average length (number of letters) of the words in the article num_links Number of links within the article num_images Number of images within the article num_videos Number of videos within the article category Indicator variables for all non Other article categories as defined by unique section tabs on mashable.com. Options were lifestyle, entertainment, business, social media, technology, and world article_subjectivity The subjectivity of the article on a scale from 0 to 1. Subjectivities closer to 0 are more factual and subjectivities closer to 1 are more opinion. title_subjectivity article_polarity The polarity of the article on a scale from 1 to 1. Positive polarity means the content of the article is associated with words considered to have a positive connotation. Values closer to 1 indicate strong positive polarity. Negative polarity means the content of the article is associated with words considered to have a negative connotation. Values closer to 1 indicate strong negative polarity article_rate_positive_words Rate of positive words in the content article_rate_negative_words Rate of negative words in the content title_polarity 1 to 1 weekend Was the article published on the weekend (1=yes, 0=no)? shares The number of times the article was shared in social media

PAGE 5

Results of Analysis Beta[1]: Constant Beta[2]: Num Words Title Beta[3]: Num Words Article Beta[4]: Average Word Length Beta[5]: Num Links Beta[6]: Num Images

PAGE 6

Results of Analysis Beta[7]: Num Videos Beta[8]: Lifestyle Category Beta[9]: Entertainment Category Beta[10]: Business Category Beta[11]: Social Media Category Beta[12]: Tech Category

PAGE 7

Results of Analysis Beta[13]: World Category Beta[14]: Article Subjectivity Beta[15]: Title Subjectivity Beta[16]: Article Polarity Beta[17]: Article Positive Word Rate Beta[18]: Article Negative Word Rate

PAGE 8

Results of Analysis Beta[19]: Title Polarity Beta[20]: Weekend

PAGE 9

Future Extension and Improvements Collect Data from wider range of sources and potentially more current data to see how this changes during the Pandemic. Fit other models to potentially improve model fit to data. Include additional attributes such as source of the articles to determine impact on share rates .