With the development and application of information technology and sensing technology, industrial manufacturing has entered the era of ‘big data’. Analyzing industrial big data and extracting valuable information has become a significant challenge in big data research. Plate shape, as a critical quality indicator in heavy plate production, is intricately linked to production process parameters and presents a high level of complexity in control. Due to the characteristics of multi-variety and small-batch production in heavy plate manufacturing, as well as the large scale and heterogeneous nature of heavy plate big data, existing single data analysis methods are inadequate for direct application. In this study, an industrial big data visual analytics approach is adopted, using data visualization as a window to integrate human perception, analysis and reasoning abilities, as well as data analysis algorithms. A hybrid model combining ‘visualization + interaction + algorithms' is proposed for heavy plate shape quality monitoring and anomaly diagnosis. A big data visual analytics system is developed to address the challenges of comprehensive analysis and utilization of big data in heavy plate shape quality monitoring. Experimental verification results demonstrate the effectiveness of the proposed approach and system.