A Data-driven Exploratory Service Composition Tool for Data Scientists

In the era of big data, data scientists gain insights from data and make decisions. However, when constructing a data analysis pipeline, data scientists are often required to be proficient in multiple algorithm models and theoretical foundations, and they have to select and combine multiple algorithm models, and repeatedly adjust algorithms and parameters to construct a high-performance data analysis pipeline. In response to the above problems, this paper proposes an exploratory service composition tool, which can perform real-time service recommendations according to users’ needs and data features, and assists users in constructing data analysis pipelines in an exploratory manner. This method can reduce the difficulty of data analysis, effectively save labor costs, and improve the quality and efficiency of data analysis.