As the coming of big data era,the need of data analysis is becoming increasingly diverse,this results in the incapability of big data analysis tools to meet the customised data analysis requirements by using its own build-in algorithm libraries,to develop or integrate new algorithm is urgently necessary. But existing big data analysis tools algorithm has high learning cost in development and integration,and makes it difficult to develop and integrate a new one. This paper proposes an approach targeted at the automatic algorithms development and integration for big data analysis tools,the algorithms are integrated into analysis tools as modules. The approach first defines the module model,and then presents the automatic generation flow of the module model,finally it puts the emphasis on analysing the automatic code generation and code detection method of modules,and proposes the metadata-based code generation scheme and the Soot control flow-based static code detection algorithm. As the experiment shows,this approach can complete the automatic development and integration for big data analysis modules.
