Madurese-Language Social Media Comment Dataset for Public Service Analysis

This dataset contains 1,510 public comments written primarily in Madurese, collected from social media platforms including TikTok, Facebook, Instagram, and X. The comments discuss various public services, government programs, infrastructure, and community issues in Madura, Indonesia. Each record includes the following information: Platform: The social media platform where the comment was posted. Kategori_Layanan: The public-service or issue category associated with the comment, such as roads and bridges, public lighting, clean water, healthcare, education, waste management, parking, flooding, public transportation, and other government-related issues. Teks Mentah: The original comment text, preserved in its collected form, including local spelling variations, informal language, abbreviations, punctuation, emojis, and code-switching. Code Mixing: A binary annotation indicating whether the comment contains code-mixing between Madurese and another language, primarily Indonesian. Sarkasme: A binary annotation indicating whether the comment contains sarcastic or ironic expression. The dataset is intended to support research in low-resource language processing, social media analysis, public-service evaluation, code-mixing detection, sarcasm identification, and the study of public opinion in Madurese-speaking communities. It may also be useful for developing and evaluating natural language processing models involving regional languages, informal user-generated content, and multilingual communication.