Pro Kabaddi League Dataset (PKL)

This dataset contains 32,341 viewer comments collected from two YouTube channels associated with the Pro Kabaddi League (PKL) — the official Pro Kabaddi League channel and the Star Sports channel — spanning six consecutive seasons (Seasons 7 through 12, approximately 2019 to 2025). The PKL, established in 2014, is one of India's most prominent professional sports leagues and is built around kabaddi, a traditional indigenous contact sport. Comments were gathered from a range of video content including live match broadcasts, recorded replays, highlight clips, short-form videos, and post-match press conferences. Raw comments — including those written in Hindi, Tamil, Telugu, Marathi, and other Indian regional languages — were translated into English using the OpenAI API, cleaned, and filtered to remove very short entries and special symbols. Each comment was then classified into one of three quality tiers: Rich (detailed, player- or tactic-specific content), Moderate (some context but limited depth), and Poor (generic or off-topic). The dataset is structured as a single CSV file with two tabs — one per channel — and each row contains the original comment, the preprocessed version, season label, video type, timestamp, like count, and classification label. The dataset is suitable for research in natural language processing, sentiment analysis, sports fan engagement, digital media studies, sports marketing, and multilingual text analytics. It is the first publicly available annotated comment corpus from the Pro Kabaddi League.