This paper proposed a visualization method for data asset value assessment based on artificial intelligence, which supported structured measurement, dynamic scoring and intuitive interpretation of data resources. Our approach fuses graph representation learning, temporal aggregation, and visual analytics engines to estimate quality level, circulation capacity, task contribution, and risk cost from metadata, usage logs, and association records. A data set covering three business domains of retail, manufacturing and financial services and containing 4860 data assets was used for evaluation. The experimental results show that the proposed method reduces the mean absolute estimation error from 0.214 to 0.087, improves the ranking consistency from 0.79 to 0.93, and reduces the batch evaluation time from 96 s to 28 s. In the interactive analysis, the response delay is kept below 420 ms, and the accuracy of abnormal asset recognition reaches 94.1%. The visualization module further supports hierarchical exploration, confidence comparison, and value formation path tracing, showing users the ability to reliably compute, clearly explain, and deploy in real-world asset valuation scenarios.
