You can become a data miner even if you only know Excel analysis

Whenever I see data miners talking about large-scale data processing and machine learning algorithms, I think this is what a data analyst should look like, this is the master.

For example, as the bureau has been managing a partner’s data mining team, whenever it is necessary to evaluate the performance of each data miner, it is found that the strongest technical ability is often ranked behind, and the highest evaluation is often the person who has an idea and is willing to cooperate, that person will not even write R or Python, but will only EXCEL analysis.

Why?

Ranking first once can be said to be accidental, but every time it is the first, there must be an inherent factor, do we really not respect technology, can technology not effectively create productivity?

Obviously not, the author feels that our cognition of the data miner has deviated, will be algorithms, will tools, will process of course is an effective means, but obviously not the only means to support the final decision, or even the most important means, this is what many people can not see, the reason why can not see, may have something to do with metaphysics, such as data mining These words themselves are from the perspective of means to describe the post, with a strong technical color, such as limiting the data, limiting the mining, Let everyone think that the more advanced professional mining methods are used in the field of data, the more likely it is to generate revenue, but the reality is much more complicated.

So, what makes a good data miner?

First of all, let’s understand what is called real professional knowledge, the author first tells a story, Cheng Jia gave such an example in the book “Study Well”:

“Xiao Ying is the editor-in-chief of Luo Ji Thinking’s “Elite Daily Class” in the App, this girl is fierce, often at three or four o’clock in the morning, if you have not “got” cooperation with Luo Ji Thinking, you will not know how crazy the people on this team work. ”

“Few people, more work, and more demanding — it seems that this is not only the pain of the design industry, but also a true portrayal of “getting” the work of the team. However, in my opinion, many unfinished tasks have been done by Xiaoying well, in her words: here, we must live alone as a team. ”

“What kind of team has Xiao Ying become as a person?” She alone is responsible for theme planning, audio recording, audio editing, content review, message review, new author mining, old author maintenance, new content development, promotional copy planning… When she threw herself into a team that could complete the abilities of “reconnaissance”, “ambush”, “sniper”, and “encirclement point” at any time, she became a slash youth. ”

Therefore, to do a good job, it is not that it is enough to learn only the knowledge of a certain profession, nor is it simply to learn this, that is also to learn, but to learn all the core abilities related to solving a certain type of problem, which must break through the professional limitations.

What we call professions, such as marketing, law, politics, history, literature, IT, or data miners, are just artificial classification labels, but the world does not operate separately within each profession according to the labels you divide. Behind a marketing problem, there are often legal, political, historical and cultural factors, but our so-called professions do not care about these: you learn the 4P (product, price, channel, promotion), market segmentation and other concepts, you can graduate, this understanding, will greatly hinder us from learning the knowledge that we really should learn.

Data miners support decision-making, so what knowledge is it about making decisions?

If you want to monetize externally based on data, you need to understand the various vertical industries, which is the most important thing in data mining, if the basic concepts of the financial field are not clear, don’t say that you do risk control models for others, one of the biggest challenges for operators to do external realization is that they do not understand the industry.

If you want to put data for precision marketing, you need to learn some marketing knowledge, so as to know the 4P, the basic concept of market segmentation, the market business process is not clear, the so-called data-driven business is also pulled, the effect of data mining is related to policies, products, channels, where is just a matter of data?

You have to understand the needs of mining you have to communicate with people, this time you have to learn some psychology, basic emotional intelligence or to have, stubbornness is the data miner to be a problem, a good miner is first of all a good listener, do data mining do not engage in any lone wolf and heroism.

If you want to show your results, you need to understand the point pyramid principle, know how to express the results of your analysis clearly and accurately, let people understand at a glance, maybe you have used the decision tree algorithm countless times, but you may not know the hierarchy analysis method.

You need to provide the data to the outside world, but also need to understand some legal knowledge, know the state’s policies on personal privacy protection and the company’s information security regulations, otherwise foolishly get the list out, make a big fuss but will be sentenced.

Data miners are often framed by secular labels to frame their own possibilities, so learning is to learn what is inside the label, read a bunch of algorithm books, learn a bunch of languages, and understand a lot of EXCEL and PPT skills, but these alone can’t actually do things.

What the author wants to say is that in this world, if you want to achieve the ultimate, you must learn “useless use”, useless use, in order to be of great use.

If we understand ability from this perspective, we have to jump out of the limitations, and we need to learn three levels of courses in our lifetime: (1) Public basic courses: executive ability; (2) Professional compulsory courses: professional ability; (3) General compulsory course: structural ability.

For data miners, the public basic course is the execution ability that each of us uses every day, such as time management, business etiquette, communication, EXCEL, PPT, mind mapping, etc. There are a large number of books on the market to introduce these knowledge, and it is more convenient for us to learn and master.

As mentioned earlier, this major does not refer to the word mining, but refers to the sum of all the abilities that can solve a decision-making problem with data end-to-end, you have to think and solve problems across disciplines, and one person lives as a team. The knowledge of this kind of systematic problem solving is often implicit, and we need to understand the subtleties of cross-domain knowledge convergence in the process of continuous practice and thinking, so as to flexibly call the knowledge between multiple disciplines at any time and win a battle, in addition to traditional data, platform and algorithm knowledge, but also include mathematical knowledge, marketing knowledge, industry knowledge, psychological knowledge, security knowledge, analysis methods and so on.

The general compulsory courses are what I saw from Chengjia’s “Good Learning”, and I think that this is some knowledge that moves towards a higher cognitive level, such as Newton’s second law F=ma, which can guide our actions more broadly and more generally, also called “critical knowledge”, Charlie Munger called “universal wisdom”, such as compound interest effect, probability theory, golden circle of mind, evolution, systematic thinking, two eight rules and so on.

For example, system thinking emphasizes “relationship”, rather than “people and things”, although data modeling is very important, but more important is the relationship, that is, the need to open up the effect data and the original model of this feedback optimization process, ChengJia also through the relationship analysis of the Wei Zexi incident, judged that the root cause of Baidu’s “evil” is the lack of a normalized feedback mechanism for the search effect, Taobao is unlikely to have such a bad problem because of buyer reviews.

For example, the two-eight principle, data mining has spent too much money on data processing, variable preparation and model release, this part takes a long time, the value is small, obviously does not meet the two-eight principle, it is necessary to reduce the length of this part as much as possible, which is why the author hopes to do some breakthrough work in agile data mining.

For data miners, being able to stand alone is the embodiment of comprehensive quality, and its level is definitely not represented by mastering several algorithms and several tools, which can explain why some people who are not very good at algorithm tools still have such strong data analysis capabilities, we often only see the “visible” professional ability, and often ignore the cultivation of “invisible” professional ability.

Be the first to comment

Leave a Reply

Your email address will not be published.


*