Evaluating the urban of low-carbon development (LCDU) is crucial for promoting sustainable urban development. Previous evaluation works are weak in exploring the influence of spatial patterns of urban agglomerations on LCDU and analyzing the differences among urban agglomerations, focusing on the selection of indicator sets rather than the establishment of a complete evaluation system. To address the lack of feature mining at the granularity of urban agglomerations in the traditional Entropy Weight-Topsis algorithm (EWT), this paper proposes the EWT deformation. Building upon this, we construct a low-carbon development evaluation system (LCDE). To demonstrate the validity of LCDE, we develop a LCDVis system with data from 62 Chinese cities as the example and design visual encoding, including high-dimensional feature portrait and feature river map to enhance the interpretability of the evaluation results. Furthermore, we take three case studies and some positive feedback from users and experts that demonstrate the effectiveness of our methodology.