Analysis and research of e-commerce user behavior data based on Spark framework

With the progress and development of The Times, data is becoming more and more important to us. The arrival of Big data has innovated technology in computers and other industries, bringing us into the era of big Data. Due to the rapid development of e-commerce, many e-commerce platforms now use big data technology or cloud computing for data management. At present, the most used big data frameworks are Hadoop and Spark. Through analyzing user behavior data obtained from e-commerce platforms, users' preferences can be guessed and recommended to meet users' needs. Therefore, understanding user behavior is a necessary condition for the development of the e-commerce industry. This paper analyzes the data from the perspective of big data. The data is processed, mined and analyzed through the real data provided by the operators of relevant platforms and corresponding results are obtained. By analyzing users' behaviors and integrating and classifying these data by means of mean clustering algorithm, naive Bayes method, decision tree algorithm and other methods, e-commerce platforms can predict users' favorite and preferred products according to these classified data, which can provide users with corresponding products in a more targeted manner and save each other's time. This paper starts with the preprocessing of e-commerce user behavior data, user behavior feature data mining and user behavior analysis, classifies user data, analyzes the data by classification, and obtains the results. Finally, through the construction of virtual machines and Spark framework environment, the user behavior data of the e-commerce platform is analyzed to obtain the corresponding user purchase information data, and the purchase intention of users is analyzed according to these data.