E-participation platforms have emerged as digital tool to facilitate citizen engagement through online deliberation, voting and oversight processes. The digital add-on for participatory democracy (Carole Pateman) can be found in different countries around the world. In Asia, the rollout of two platforms in Taiwan, namely iVoting and Join, has captured the attention of Western media outlets. However, there is little literature on the exact content debated and agreed on these platforms up to this point. Utilising recent advancements in NLP, we explore the content of the proposals that were made. In our study, we combine new approaches of text mining with political analysis on Taiwan’s e-participation platforms. The dataset, which includes 14,118 proposals from 2015 to 2022, has resulted in a distinct topic model being constructed for each platform. With the help of our method, we were able to cluster the proposals thematically and show which concerns were articulated and with how much approval. Based on a random sampling of 110 proposals, we were able to determine that our method assigns 81.82% of the proposals to the corresponding cluster. This can also significantly overcome language barriers, as we employed a translation pipeline within the text-mining process from Chinese into English. Our method is adaptable to e-participation platforms in various languages, providing decision-makers with a more comprehensive tool to understand citizens’ needs and enabling the formulation of more informed and effective policies.