摘自数据分析入门
1.Question Phase. The process begins with the question you want to answer or a problem you want to solve.
（1）未通过项目的学生有什么特点?
（2）怎么根据顾客需求存货？
2.Wrangling Phase. 数据采集和数据清理。
3.Exploration Phase. Get familiar with the data,building the intuition and finding patterns.
4.Draw conclusion Phase. draw the conclusion or make some prediction.
eg. predict : which movies users will be like?
conclusion:users are less likely to click certain articles.
5.communicaiton phase. your findings are only as useful as your ablity to communicate them.even the system,you also need to share what you have bulit and how it works with your team.
communication :
1. post a blog
2. a paper
3. an email
4. power point presentation
5.face to face communication
Data visualization is almost always useful
通常 data wraggling and data exploration 是交替进行的。 首先数据清洗，然后探索，然后进一步清洗数据，进一步探索。