This is a full-stack news mining and recommendation web application. It uses React for the front-end, Node.js for the web service and Service-oriented pattern of Python backend. I applied the frequently used Python libraries, deployed and connected MongoDB and RabbitMQ, and utilized web crawling to obtain news. At the meantime, I implemented TF-IDF to remove duplications over texts and CNN to realize news recommendation. At the end, I used TensorFlow to realize the News Topic Modeling, and deployed TensorFlow Serving to provide online prediction.

1. [] Understood the workflow and difficult points of web crawling and page data extraction, and learned how to use web crawling to obtain the information I actually needed.
2. [] Understood and mastered the React framework, and knew how to build a site using React front-end framework.
3. [] Became familiar with the Service Oriented Architecture and knew how to use this architecture to design a system.
4. [] Became familiar with the use of Message Queue, learned how to deploy and connect RabbitMQ, and used RabbitMQ for load balancing.
5. [] Learned the basic knowledge on Text Mining, using TF-IDF for text de-duplication.
6. [] Understood the fundamental theory and workflow of Machine Learning, became familiar with several common machine learning problems and models, mastered the basic Deep Learning skills and the use of TensorFlow, and learned how to use CNN to implement a recommendation system.
7. [] Mastered the Python libraries, and learned how to use RPC to build a Python backend.