The proliferation of rich social data sources in the big data era offers a valuable opportunity for studying the planning of art events. In this study, we harvested data on art activity planning from social media via microblogging API calls, converting and analyzing this data through a novel methodological framework. Specifically, we employed the Sentiment-Enhanced Deep Graph Convolutional Network (SEDGCN) model to extract and identify sentiment features associated with art activity planning from social media datasets. Sentiments were categorized using the softmax function. Subsequently, our analysis integrated these findings into the art activity planning process. By conducting a correlation analysis between positive and negative sentiments on social media and various elements of art activity planning, we found a significant correlation (p < 0.005) with all six examined elements of art planning. Furthermore, an effectiveness analysis conducted post-implementation of the planned art events revealed predominantly positive emotional responses among attendees. Notably, 325 art professionals reported a sense of healing as a result of their participation. The methodology proposed in this paper for analyzing social media data effectively captures audience emotions, thereby assisting planners in crafting art events that resonate with and fulfill the emotional needs of the audience.