CryptoVision: A Comprehensive Dataset for Crypto News and Trend Prediction

This dataset presents a large-scale collection of 188,430 cryptocurrency news records published between 2017 and 2025. The data were gathered from six reputable and verified sources — BlockWorks, CoinDesk, CoinTelegraph, CryptoPanic, CryptoNews, and Decrypt — using Python-based automated collection pipelines. Each record provides rich and structured metadata about a cryptocurrency-related news article, including the title, publication date and time, cryptocurrency type, and source URL. No copyrighted article text is included; only publicly accessible metadata are shared to ensure full copyright compliance.

To align market behavior with news events, each article is matched with Binance OHLCV candle data of a 15-minute timespan, allowing researchers to study short-term price reactions following news releases.

Each record in the dataset consists of the following columns:
URL, Title, Date_Time, Coin_Type, Sentiment_Label, Sentiment_Score, Open, High, Low, Close, Volume, Movement_OpenClose_%, Movement_HighLow_%, and Market_Move.

This dataset enables comprehensive sentiment analysis, market impact assessment, and event-driven modeling in cryptocurrency research, offering a unique link between news sentiment and real-time price fluctuations.