Subject Guides: Social Media Research: TikTok

TikTok TOS and API

As with many social media platforms, the public TikTok API provides limited access and the TOS bans all automated scraping. However, the TikTok API is even more restricted than the APIs offered by Meta. You may notice that the TikTok Developers page only offers the options to use TikTok as a sign-in option on your site, post videos via other apps, share music to TikTok's library, and embed TikTok videos on other platforms. None of these options are helpful for research uses.

Additionally, the mostly-video content of TikTok poses a further challenge to collection and analysis of datasets compared to text-heavy platforms like Twitter or Reddit. These factors, combined with the relatively recent arrive of the platform on the social media juggernaut scene, mean studies of TikTok have been difficult and limited.

The TikTok client uses a sophisticated system of signed requests to prevent third-party systems from making requests of TikTok servers. This is part of TikTok's broad efforts to track users, which also includes IP address tracking, browser fingerprinting, and tracking cookies. A few web scrapers or unofficial APIs, which query the private internal TikTok API by simulating a web browser to circumvent the signed request challenge, can be found. Some are quite robust, but require setup and/or coding knowledge and violation of the TikTok TOS. TikTok will also attempt to block your connections as soon as it detects scraping. The report from Mozilla linked at the bottom of this page details some of the difficulties of automatic data collection.

Tools for TikTokk Research

TikTok, as the newest platform on this list, has not had the benefit of a decade or more of academic study and the development of research tools. Additionally, with virtually no useful public API, automated data collection necessitates violating the TOS with web scraping or complex browser-simulating reverse-engineered calls to TikTok's private API.

Like Instagram, at this time (05/2022) any tools for automated extraction of videos, or text accompanying videos, are likely to violate TOS. Additionally, the ones that do exist require at least some coding or command line knowledge. The automated tools on this list - drawrowfly's TikTok scraper and the Bellingcat Hashtag Analysis Toolset built on top of it - are provided as an example of this sort of tool, but they are complex, risk a ban from TikTok, and require coding or command line knowledge. Researchers are encouraged to explore manual collection methods.

Bellingcat: Investigate TikTok Like A Pro!
A guide from open source intelligence journalism outlet Bellingcat on manual data collection and search techniques on TikTok.
Zeeschuimer
A Firefox browser extension that helps with manual data collection by recording information from posts on Instagram and Tiktok as you scroll through them. Exports as a .json file or directly to 4CAT. Captures a large amount of metadata and will need cleaning or loading into 4CAT to be easily readable. Does not capture images or videos. Does not violate the TOS.
4CAT
4CAT is a relatively advanced tool for the collection and analysis of social media data - it's best run on a UNIX server and has dependencies that it does not automatically install itself - but with the upside that it has modules built to work with important but niche platforms like 4chan, 8kun, Parler, and more, as well as Twitter and Reddit. Works with TikTok metadata collected via Zeeschuimer.
Bellingcat's TikTok Timestamp date extractor
Very simple site that decodes Tiktok video urls to get the actual creation date, which is not otherwise shown. Does not violate TOS.
drawrowfly's TikTok Scraper
A powerful node.js application that uses unofficial API calls as well as web scraping to download videos, post text, and user/video/hashtag metadata. Violates the TOS!
Bellingcat's TikTok Hashtag Analysis Toolset
Builds on drawrowfly's node.js application to help identify co-occurances of hashtags. Link points to post on Bellingcat site giving examples of the toolset's use.

Example publications in TikTok reserach

‘#OkBoomer, time to meet the Zoomers’: studying the memefication of intergenerational politics on TikTok
ABSTRACT: TikTok, a short video platform featuring content between 15 and 60 seconds long, has become a popular and rapidly growing social media application around the world. As a platform catering for light entertainment, TikTok champions virality and encourages memetic remixes. Meme videos, mostly featuring lip-syncs, dance routines, and skits, have become one of the defining features of the platform. These seemingly trivial videos have been utilised by young TikTokers to advocate for various causes.
more...less...
This paper uses #OkBoomer memes as a case study to examine the political culture of young people and Gen Z in particular. By analysing how intergenerational politics has been ‘memefied’, this study delineates how Gen Z imagines and expresses a generational sentiment towards ‘Boomers’ as the imagined other. They do so through short video cultures and practices on TikTok, drawing upon the networked experiences of their peers. Specifically, the paper considers the key controversial issues, meme forms and meme functions across the #OkBoomer memes on TikTok, and its eventual mainstreaming in society.
Through a white lens: Black victimhood, visibility, and whiteness in the Black Lives Matter movement on TikTok
In this paper, we explore how highly visible users in the context of #BlackLivesMatter on TikTok shape the narrative around Black victims of police brutality, the understanding of these narratives by others, and the potential consequences of these portrayals for the movement at large. To examine these dimensions, we analysed the 100 most circulated TikTok videos and associated comments depicting victims of police brutality using the #BlackLivesMatter hashtag through multimodal critical discourse analysis.
more...less...
We identified how users attempted to increase visibility of their content, and how this was supported or criticised by commenters depending on the perceived motives of these efforts. Furthermore, we showcased how influencers raised awareness of the movement with little personal effort or risk, sometimes appearing to leverage the movement for self-exposure. Our analysis showed that many of the most liked videos were made by white content creators who, in their videos, seemed to be addressing an imagined white audience. While these efforts portrayed the movement favourably, the content creators remain outsiders who have not themselves been in harm's way of police brutality. While there were exceptions that promoted the perspectives of marginalised communities, and while the white narratives were consistently supportive of the movement, they also work to displace focus on racial (in)justice away from those directly affected by it, that is, away from Black people’s own experiences of police brutality. We discuss these findings in relation to questions about digital representations of Black victimhood, digital visibility and practices of whiteness, on TikTok and beyond.
Strategies for Studying TikTok During the German Election
A report from Mozilla detailing the methods underlying their investigation of political disinformation on TikTok prior to the 2021 German Federal election.