Abstract: Online judge submission log refers to the records generated by users submitting program through the online evaluation system,including user,code,submission status and other information,which reflects user’s online programming behavior.While existing online judge system usually provides users with topic list and real-time ranking information,fails to provide visual information on online programming behavior of users from such dimensions as timing sequence,topic type and error type,so that users cannot evaluate their strengths and weaknesses of programming ability in time.In order to help users evaluate their programming ability and improve their learning efficiency,based on the muti-dimensional quantitative analysis of user’s ability to solve problems in basic algorithm,data structure,search,graph theory,dynamic planning,mathematics,etc.,the computer program radar of user’s programming ability is constructed to realize muti-dimensional visualization of user’s programming ability.According to the model,a muti-dimensional visualization of single user and group user’s programming behavior for online submission log is designed and implemented.Finally,by analyzing the submission log of POJ of Peking University,our method realizes the real-time evaluation of single-user muti-dimensional programming ability and the visualization analysis of system group user programming behavior,and provides support for collaborative filtering and recommendation of programming problems.