Cryptocurrencies, such as Bitcoin, are becoming increasingly popular, having\nbeen widely used as an exchange medium in areas such as financial transaction\nand asset transfer verification. However, there has been a lack of solutions\nthat can support real-time price prediction to cope with high currency\nvolatility, handle massive heterogeneous data volumes, including social media\nsentiments, while supporting fault tolerance and persistence in real time, and\nprovide real-time adaptation of learning algorithms to cope with new price and\nsentiment data. In this paper we introduce KryptoOracle, a novel real-time and\nadaptive cryptocurrency price prediction platform based on Twitter sentiments.\nThe integrative and modular platform is based on (i) a Spark-based architecture\nwhich handles the large volume of incoming data in a persistent and fault\ntolerant way; (ii) an approach that supports sentiment analysis which can\nrespond to large amounts of natural language processing queries in real time;\nand (iii) a predictive method grounded on online learning in which a model\nadapts its weights to cope with new prices and sentiments. Besides providing an\narchitectural design, the paper also describes the KryptoOracle platform\nimplementation and experimental evaluation. Overall, the proposed platform can\nhelp accelerate decision-making, uncover new opportunities and provide more\ntimely insights based on the available and ever-larger financial data volume\nand variety.\n