To conduct an in-depth analysis of the causes of civil aviation accidents and prevent accidents, given the incompleteness of current accident data statistics and the need to make accident causation analysis credible and profound, accident report data were collected as samples from global civil aviation accidents in the year 2000. These data were subjected to structured processing, including the categorization of accidents into stages and types. Using the K-means algorithm, a cluster analysis of civil aviation safety risk factors was performed from multiple perspectives. A comparison was made between the results of structured and unstructured data processing. Various risk factors" combined impact on accidents was considered. Key risk factors identified from word clouds and latent topic distribution in text data using LDA topic modeling were analyzed. This comprehensive approach provides valuable reference information and safety recommendations for future civil aviation safety.