Mapping the Climate Change Landscape on TikTok

TL;DR

The paper addresses how climate change discourse unfolds on TikTok, a platform with a young audience. It builds a public dataset of 590,361 climate-related videos from 14k creators and their follower network, and constructs a data-driven taxonomy of climate topics; it then uses BERTopic and LLM labeling to derive 26 topics (17 climate) and analyzes topic and user networks to identify gateway non-climate topics, notably travel and political content, that can recruit new audiences into climate discussions. The findings show concentration on Lifestyle, Diet, and Nature topics and reveal meaningful cross-topic links that could inform climate communication strategies on short-form video platforms. The work offers practical guidance for designing campaigns to broaden climate awareness among youth and outlines avenues for future research on influencer coordination and endorsement effects.

Abstract

Social media platforms shape climate action discourse. Mapping these online conversations is essential for effective communication strategies. TikTok's climate discussions are particularly relevant given its young, climate-concerned audience. In this work, we collect the first TikTok dataset on climate topics. We collected 590K videos from 14K creators along with their follower networks. By applying topic modeling to the video descriptions, we map the topics discussed on the platform on a climate taxonomy that we construct by consolidating existing categorizations. Results show TikTok creators primarily approach climate through the angle of lifestyle and dietary choices. By examining semantic connections between topics, we identified non-climate "gateway" topics that could draw new audiences into climate discussions.

Figures (3)

• Figure 1: Taxonomy of climate-related topics. Labels are color-coded based on their occurrence: purple indicates presence only in previous works, orange shows topics found exclusively in our data collection, and magenta represents topics present in both. Non-climate topics are shown in blue. Superscript numbers on purple and magenta topics indicate the work where these topics were originally identified.
• Figure 2: (a) Projection of the user-topic bipartite network. Topic $T_1$ is discussed with high frequency by 3 users ($v,w,z$), while topic $T_2$ by $4$ users ($w,x,y,z$). In the projected graph on the topic layer, the edge $(T_1, T_2)$ has $w = \frac{2}{5}$. (b) Climate and non-climate bridge topic pairs. Non-climate user $A$ has received $10$ likes on topic $T_1$. This is divided by its 3 neighbors and weighted by $40\%$ as it represents the strength of topic $C_1$ for climate-user $B$. So, the weight of the edge $(T_1, C_1)$ concerning the couple of users $(A, B)$ is $\frac{10\cdot 0.4}{3}$.
• Figure 3: Weights on topic pairs. (a) Topic perspective. The heatmap reflects the proportion of shared userbase. (b) User perspective. Rows correspond to non-climate topics while columns collect climate-related topics. The weights represent the potential visibility that non-climate topics could bring to climate-related topics.