The product represented by this dataset is manganese concentrate, a high-grade manganese ore obtained through beneficiation enrichment, in which the manganese content is significantly increased to the level required for metallurgical or chemical utilization. Its common forms are powder or granules, with a density of about 3.7-4.5 g/cm³. The specific grade varies with the intended use: the metallurgical grade is mostly 35%-48%, while the chemical grade can reach higher purity. The main impurities include iron, silicon, phosphorus, and aluminum, which affect the efficiency of subsequent smelting or chemical processes and product quality. The upstream key raw material of this product is low-grade primary manganese ore, from which the concentrate is obtained through crushing, grinding, and beneficiation enrichment. Downstream, it mainly serves as an important raw material for smelting manganese-series alloys (silicomanganese, ferromanganese, electrolytic manganese), and is also used to produce chemicals such as potassium permanganate and manganese dioxide, as well as refractory materials. The system boundary includes sub-processes such as raw-ore pretreatment, grinding, beneficiation, dewatering and drying, and the treatment of waste gas, wastewater, and solid waste, explicitly excluding mine exploitation, long-distance transport of raw materials, manufacturing of production equipment, and plant infrastructure construction. The data sources integrate industry-average indicators, on-site measured data from typical beneficiation plants, and relevant beneficiation process literature, and are constructed by merging multiple sources. Where the production process involves multiple-product situations, mass allocation is carried out according to the product mass ratio to ensure that inputs and outputs correspond to the actual process flow. The dataset includes the main raw and auxiliary materials, energy, and waste in the production process; infrastructure such as the manufacturing of production equipment and the construction of plants is not included within the dataset boundary. The data inventory is formed by merging data from multiple sources.