Flink+ClickHouse builds a hundred million-level e-commerce user portrait platform (PC, mobile, applet)
Full version 131 sessions, new class in October 2020
This course adopts the Flink+ClickHouse technical architecture to realize our portrait system. After completing this course, you can save your exploration time, save corporate costs, and improve corporate development efficiency.
I hope this course will be helpful to some enterprise developers and partners who are interested in the new technology stack. If you have suggestions for the content of the tutorial I recorded, please communicate in time. The algorithms used in the project include Logistic Regression, Kmeans, TF-IDF, etc. Flink temporarily supports fewer algorithms. For the above algorithms, this course will take everyone to implement them with Flink, and combine the real scenes, and use them after learning.
The system includes all terminal data (mobile terminal, PC terminal, applet terminal), supports analysis and query of hundreds of millions of data, and performs portrait calculation of users in real-time and near real-time. The profile indicators included in this course include: overview trends, basic attributes, behavioral characteristics, hobbies, risk characteristics, consumption characteristics, marketing sensitivity, user tag information, user groups, product keywords and other major indicator modules, each indicator Will take everyone to realize it by hand.