The exponential growth of academic information on the Internet has created a pressing need for efficient methods to query, analyze, and reveal its potential value. However, the challenges in collecting, organizing and effectively utilizing the data remain significant obstacles. We construct CoopKG, a new academic knowledge graph (KG) that integrates extensive data on academic papers, researchers, and research projects. CoopKG is stored in a Neo4j database and is linked to comprehensive profile documents and portrait images of researchers. To fully leverage CoopKG, we develop a Knowledge Graph Question Answering (KGQA) system that utilizes innovative techniques with large language models (LLMs) to accurately convert a natural language question into Cypher Query Language (Text-to-CQL) for retrieving relevant information, thereby generating highly readable answers. Additionally, we integrate extensions such as graph visualization, data export, online search, and AI-assisted reading to enhance the system’s usability. We conduct extensive experiments on open-source LLMs, demonstrating that on the Text-to-CQL task, our method can improve the logical accuracy by up to 5.71% and the execution accuracy by up to 5.29% on fine-tuned models.