Data science has become a basic tool to extract value from data. Any enterprise can take data collection, storage and processing as a part of its business. The huge multi-source heterogeneous data flow has also become an important mysterious asset to be deeply excavated because of its rich information. Based on the basic models and methods of business intelligence and data analysis introduced in the course of “Introduction to Data Science” (i.e. the "Meta-method" of Data Science), this course will deeply explore the practical application scenarios and cases of data science methods and models in different fields (i.e. the "combination algorithm" of data science, such as recommendation system, sentiment analysis, search engine principle, fraud detection, customer segmentation, relationship marketing, etc.). By using Rapidminer, Python and other platforms and tools, we can enable students to strengthen their ability to practice and solve data science application problems while learning and mastering theoretical methods.