Epidemiological modelling of online social network dynamics (John Cannarella & Joshua A. Spechler (2014).
My view on the research paper.
1.) Assumptions of the model and the data analysis used in the research work?
-An assumption is that users typically join Online Social Networks because their friends have already joined. This might the true, but certainly not the case for every individual.
-Another assumption was that the population remains constant during the outbreak.
-Also, the assumption that Google search query data is the only means for accessing OSNs, and that it also represents the magnitude of the infected population of the model.
2.) Some arguments that could call the authors’ conclusions into question?
-Google only track trends if we assume everyone accesses OSNs through the search engine. As data can prove that this isn’t the case, the majority of OSN users would prefer to use the designated applications downloaded from App stores or other sites. This calls the authors’ conclusions into question.
-OSNs constantly change their models by bringing new services and products that appeal to other users. Facebook has over the years introduced products like Fb messenger, Marketplace, Live videos, and also made some acquisition which made it retain existing users and gained even more.
-Legal issues, censorship and bans are a big factor when considering studies such as this. Facebooks’ censorship in some countries (like China, NorthKorea etc) can contribute to a possible rapid decline/adoption in user base. This was not accounted for in this study.
3.) Check the statistics of the keyword “Facebook” in Google Trends. Compare it with the prediction proposed in the paper, as well as with data search for Twitter, Youtube and Whatsapp. Comment the results!
Since the paper was published in 2014, I checked to see the trends of Facebook from 2013, there is a decline as seen from the graph, the decline is not up to 80% as predicted, and the prediction of OSN adoption or abandonment rate cannot be evaluated only from google trends.