Although for a long time it was considered an entertainment platform, its enormous popularity and the arrival of other actors, such as politicians and the media, have positioned TikTok as a source of information and a space for public debate. 

While in the United States 10% of the adult population is informed through TikTok, in Colombia the consolidation of this social network as a political scenario is undeniable, as was demonstrated in this year's presidential campaign. Rodolfo Hernández, one of the least favored candidates at the beginning of the campaign, supported his communications strategy on TikTok and managed to reach the second round. In turn, his success launched other candidates into the world of short videos and trends, sharing content that at times blurred the line between entertainment and political campaigning.

In addition to occupying an increasingly large space in the digital ecosystem - with more than one billion monthly active users - TikTok has become a place for discussion of issues of public interest, promotion of social agendas, coordination and activism.

However, its growth has not kept pace with network research, as its dynamism and characteristics make it difficult to establish a methodology analogous to those traditionally used to explore social networks. 

At Linterna we analyze how public opinion is built online from the study of large amounts of data. For this reason, we have put the magnifying glass on TikTok as a robust source of information to analyze digital public debate. In this blog, we expose some problems to face when doing research on TikTok. 

‍

Lack of access to information 

Digital research usually uses alphanumeric data, interactions and trends as input. Unlike other networks, such as Twitter or Facebook, where it is possible to access information about the content, TikTok does not have an API (programming interface) to access this data and download it in large quantities for analysis. This has made it necessary to go back to the creation of databases manually by researchers, requiring a greater amount of time and limiting the collection capacity. 

In addition, TikTok is a network that promotes the creativity of its users when creating content. The editing tools, as well as the audio and music banks facilitate the creation of videos that become trends using other publications within the same platform. This self-referential character facilitates the creation of content that is not always understandable outside the borders of the platform. In terms of the research, this is an extra obstacle when it comes to constructing classifications or generalizations that allow us to generally characterize the conversations on TikTok. 

‍

New observation challenges 

Digital research in social networks has succeeded in developing a long-range metadata reading based on content available on the network. These analyses make it possible to understand the dynamics of a conversation over extended periods of time. However, TikTok has broken this paradigm and has imposed the need for new observation models. 

Traditionally, digital research has been based on data mining from written text. Therefore, approaching sources in video format, as in the case of TikTok, is a new paradigm that raises questions about how to analyze a large amount of video pieces, without having to cover them individually and manually. Especially because this task has been advanced through complex recognition algorithms that require advanced processing and out of reach for most digital researchers. 

On the one hand, the content within the platform is not fixed, every day large amounts of videos are uploaded and, transversely, many others are removed. This was the case of academic Adina Giotomer, who, upon undertaking research on this social network, noticed that a large part of her sample had been removed from the platform. "Tiktoks are ephemeral," she concluded. 

On the other hand, digital investigations are often based on text analysis, while TikTok does not allow any format other than video. This makes it necessary to rethink the whole quorum of data to be extracted in the middle of the research.  

‍

Changing and unexplored format

TikTok is constantly being updated. From time to time, functions and tools appear and disappear within the application. This changing nature is reflected in the instability of the sections, the design and the affordances -understood as the different purposes that users can give to the new functions of a platform-. These changes, which include the increase in the length of videos or the creation of new sections in the feed, directly affect the analysis, as they remove or add important sources of information.

Structural changes transform the nature of content and interactions, create new trends and cause traffic within the platform to diversify. In this way, the reading of data is hindered, as it is no longer clear what the metrics of likes, comments and the option to share videos tell us. 

On the other hand, since it is a network that functions as a recommendation system, i.e., in which the interaction of users with the content is more important than with other users, the way in which publications are presented to those who use the application is of special interest to researchers. 

‍

Given the growth of TikTok and the place it is beginning to take in the construction of public opinion, it is necessary to start tackling head-on the problems of researching this particularly broad, diverse and changing part of the digital environment. The methodological challenges are still waiting to be solved and it is essential to prepare ourselves to analyze the mine of information that TikTok has become.

‍