Online transactions platforms such as e-Commerce and social networks generate massive volumes of data, which are complex and with sizes that can reach terabytes ( \(10^{12}\) bytes) or petabytes ( \(10^{15}\) ), making data mining and recommendation tasks increasingly challenging. Existing systems for high utility sequential pattern mining (HUSPM) extract valuable (e.g., profitable) e-commerce sequential products to recommend to buyers considering both frequency of items and utility (e.g., profit, importance) of itemsets. Existing HUSPM systems include those named as HUSREC21, HUSP21, and HUSP-SP23, mine high utility sequential patterns using a single machine for these HUSPM tasks, resulting in inefficient processing of big sized datasets, longer execution times and high memory consumption. This paper proposes a system called Big High Utility Sequential Pattern Recommendation System (BIGHUSREC), that extends the HUSREC21 system to mine high-utility sequential patterns from big sized datasets through a "Top-K" approach integrated with the MapReduce framework. The focus is on extracting the Top-K most valuable (profitable) and yet relevant to the user patterns, with the aim of improving recommendation accuracy, while minimizing execution time. The proposed system uses MapReduce to partition data into smaller parts (Mapping), and analyzing each part in parallel to identify profitable patterns before aggregating the results (Reducing) for a comprehensive output. By combining purchase and clickstream data, the proposed BIGHUSREC effectively improves recommendation accuracy.