Twitter APIs for Collecting Data of Influenza Viruses, A Systematic Review

Abstract

Twitter has become famous due to its lenient policy for providing data to a variety of beneficiaries, as the retrieved data will be used according to the need and method chosen by the developer or researcher to accomplish their work. Unlike many social media networks, Twitter overcame limitations within its network and applied extra features making the usage of this platform much fun and determine. This paper will not explain Twitter's specifications and characteristics or usage instead, it will focus on describing the methods of aggregating data and demonstrate the main challenges during collecting the huge amount of data available on this platform. The Twitter Application Programming Interface (APIs) will be discussed, concentrating on the most common ones, especially those concerned with scraping and crawling Twitter data. These APIs are bounded and need solutions due to the various kinds of retrievals they make, yet, this paper will explain them and discuss appropriate matters to bypass these constraints. Furthermore, this review focuses on reviewing studies and researches collected from various journals using Twitter data to predict, detect and monitor the outbreak of influenza demonstrating all types of viruses caused it, stating the location of such spread. These flu viruses have been appeared since 2009 starting with Swin flu then H1N1, Avian flu, Flu, lastly and, still exist Covide-19, being discussed on Twitter platform, taking into consideration the launch of Twitter platform was in 2006. The importance of this review is that it shows the gradual emergence of the different dangerous flu viruses spread around the world, showing the history of two things; the appearance of the dangerous flu viruses (in the form of their collected datasets) and the developments and progresses applied to the APIs used for collecting these big datasets from Twitter platform, limitations and, constraints within these APIs are quite different during this period. Several tables abbreviate the most important information gained stating which method is the best and how advanced this process is over the last decade. The final result indicates the increasing number of researches recently and the use of very sophisticated methods to collect tweets, as a result of the emergence of the Coronavirus.