Generally speaking, I have been in formal contact with data analysis for almost a year, and I still have some experience in crash. A good data analyst can't be fast, but there are shortcuts to zero experience. The above premise is aimed at getting started, the purpose is to reach the threshold of data analyst and get the offer smoothly, without involving advanced skills such as data mining. My method is biased towards the internet field, which is suitable for both the position of analyst and the ability training of operation and products. In other fields, people have different views. There are seven-week database and seven-week programming language on the market. Today, we will learn how to become a data analyst in seven weeks and seven weeks. If Excel plays smoothly, you can skip this week. However, since I can't vlookup when I enter the business, it is necessary to mention it. The focus is on understanding various functions, including but not limited to sum, count, sumif, countif, find, if, left/right, time conversion, etc. You don't need to learn all Excel functions, it's important to learn to search. That is, how to describe the problems encountered clearly on the search engine. I think it is enough to master vlookup and PivotTable, which are the two most cost-effective skills. Vlookup is easy to understand, join in SQL and merge in Python. Learning pivot table, group in SQL is the same as pivot_table in Python. If these two points are achieved, it is not difficult to count the data within 65438+ million, and 80% of office white-collar workers can be killed. There is a classic saying in the field of data visualization and data analysis that words are not as good as tables, and tables are not like pictures. Data visualization is one of the main directions of data analysis. In addition to advanced analysis such as data mining, many data analysis are monitoring data and observation data. Training of analytical thinking. There is no problem for database learning Excel to process data within100000, but there is no shortage of data in the Internet industry. But whenever the product has a little scale, the data is millions. At this time, you need to learn the database. You can understand the principle of MapReduce. Statistical knowledge learning, business understanding is more important than data methodology for data analysts. Of course, it is a pity that there is no shortcut to business study.