Taobao training, Amoy blue ocean to teach you to learn data analysis
data analysis, the first to be clear is the purpose, no purpose, no direction. The purpose is to do a data analysis of the main line, according to the purpose of what data to go, to go through what treatment, then what kind of analysis, what kind of report.
so, we have to first step of data analysis, we first assume a proposition, that is our purpose, do business people know, for flow shop, is a very important thing, but to have traffic, to show, show a lot of points, various kinds of activities. Train drill, of course we most love the free traffic, free show. The core is still, want to have a show, you have to rank. Then we might as well look at the rankings in the end what kind of rules.
an article, it is necessary to talk about such a big proposition, so I give you a demonstration is part of it. It is well known that in the free flow, comprehensive entrance is the largest, and the effect of comprehensive entrance a big factor is for the influence of the ranking time frame (except 3 chunks, then this is the other) to see on the shelves of time points for ranking that its influence what kind of.
currently has a purpose, we must choose the data, we need what data to analyze, the first can think of, is bound to a certain point in time, in one of the following keywords, ranking the first few pages of the specific time frame under the baby.
search keywords you want to know the baby, pull out the relevant data (I use Yu Bao box pull data, you can also use other software acquisition):
here we have the time we need to go up and down, this time you have to do a good job record, what keywords, the number of words is the number of baby, what time to pull down.
I’ll do a record. This record is easy to view later, including the possibility of further data or other research.
well, we completed second steps, we have the data, then, is to test some of the basic skills of the tools we use, we have to deal with the entire data. Light so look at the data, may see some things, but not necessarily rigorous, chance is relatively large, and we need to analyze the dimension of time, time here is not a sequence that is a variable, but this is not a numerical variable type, cannot be directly used to make a direct comparison. And sometimes do not deal directly to see tens of thousands of data, see the head will be large. So it’s important to deal with the data and clean it up.
so here I am