Search on TFTC
New Study Sheds Light on TikTok's Algorithm Through API

New Study Sheds Light on TikTok's Algorithm Through API

Apr 14, 2024

New Study Sheds Light on TikTok's Algorithm Through API

In a recent study spearheaded by researchers Francesco Corso, Francesco Pierri, and Gianmarco De Francisci Morales, the veil has been lifted on one of the world's most explosively popular social media platforms, TikTok. Known for its short-form, algorithm-driven content, TikTok has emerged as a digital juggernaut, especially among younger demographics. The study, which draws data through TikTok's newly released Research API, offers a rare window into the patterns and behaviors on the platform.

Understanding the Research API

The Research API is a tool that allows researchers to gather data on user activity, video posts, and associated comments. The release of this free resource aligns TikTok with platforms like Facebook and Twitter, which have previously granted data access to researchers, especially under regulatory pressures like the EU Digital Services Act. TikTok's API, however, stands out as a relatively unexplored resource, largely due to its recent introduction and the platform's brief history.

Sampling TikTok's Universe

The study conducted by Corso and colleagues hinged on evaluating the API's reliability. They collected and scrutinized over half a million randomly sampled TikTok videos spanning six years, from January 2018 to December 2023. Their analysis probed into video engagement metrics—likes, shares, comments, and views—exposing the platform's temporal growth and the geographic spread of its user base, which is largely dominated by Asian countries, with the United States as a notable Western outlier.

Virality and Geographical Distribution

One of the study's key revelations was the significant impact of viral hashtags on user engagement. Hashtags associated with TikTok's "For You" feature, which drives content recommendations, were found to markedly increase video views and likes. However, this effect showed signs of diminishing over recent years, suggesting potential adjustments in TikTok's recommendation algorithms to balance content visibility.

From a global perspective, the study noted a heavy concentration of content originating from Asian countries, with India leading despite a 2020 ban on the app.

The Temporal Rhythms

In examining when most TikTok content is posted, the study noted peculiar patterns. For instance, an unusually high volume of videos appeared on the first day of each month, and Saturdays emerged as the most popular day for uploads. These anomalies hint at possible quirks in TikTok's internal systems rather than user behavior, an essential consideration for those studying social media trends.

Conspiracy Theories on TikTok

The researchers also touched upon the presence of 'conspiracy-related content.' By filtering through hashtags associated with known conspiracy theories, the study provided an estimate of such content's prevalence on TikTok. While the absolute numbers seemed low, the sheer scale of TikTok's user base implies that the actual count of conspiracy theory-related videos could be in the hundreds of thousands.

Conclusion and Implications

The study's insights are largely significant for future researchers aiming to understand the digital ecosystem of TikTok. It highlights the limitations and biases inherent in the Research API, and the importance of accounting for these when drawing conclusions about user behavior on the platform.

Original Study


Current Block Height

Current Mempool Size

Current Difficulty