Abstract:According to the demand of water conservancy public opinion and business work for Internet public opinion big data, based on multi-modal intelligent analysis architecture, a water conservancy public opinion big data analysis platform is constructed by using big data, natural language processing and other technologies, using knowledge atlas, information extraction and value identification to find, index and extract information value points, sensitive points, key points and fermentation points. The operation of the platform effectively promotes the integration of public opinion work and water conservancy business work, expands the depth and breadth of water conservancy public opinion work, verify the feasibility of the water conservancy business system using the Internet dimension data approach, open the development space for the superposition of Internet dimension data and water conservancy business data, improve the value density of public opinion big data, which could provide business data support and analysis decision support for business systems such as the portal website of the Ministry of Water Resources, river and lake chief system management, water resources management and water resources big data analysis.