Summarize 8 kinds of
data analysis thinking, and use some short stories to illustrate.
1. Contrastive Thinking
In our daily work and life, contrast thinking can be seen everywhere.
For example, Xiao Ming had a bad grade in a final exam and only scored 30 points in English. Xiao Ming's mother said to him: "You scored 70 points in the English test last time. Why did you get so bad this time? My classmates all scored above 80 this time."
From this example, it can be seen that there are usually two directions for comparison. One longitudinal direction refers to the comparison at different times, such as comparing Xiao Ming’s last exam results with this time. One is horizontal, which refers to comparing with the same class, such as comparing Xiao Ming's classmates.
2. Subdivision thinking
Subdivision can be said to be everywhere, as large as the universe can be subdivided, and as small as the nucleus can also be subdivided. The big goal of life can be subdivided, and the results of a small test can also be subdivided.
For example, Xiao Ming's overall score in a certain exam was not good. When I looked at it in detail, I found that his scores in other subjects were good. Only his English score was particularly poor, with only 30 points, which lowered the overall score.
This example is to subdivide the overall test scores into specific subjects. In the work of
data analysis, the subdivided latitudes mainly include time, region, channel, product, employee, customer, etc. DuPont analysis method and McKinsey's MECE analysis method are essentially subdivision thinking.
3. Traceability thinking
Sometimes, even though comparative thinking and subdivision thinking are used, no conclusion can be drawn from analysis. What should I do?
At this point, you can try traceability thinking, trace the detailed records of the data source, and then think about the logical relationship that may be hidden behind the data source based on this, and there may be unexpected insights.
For example, through comparative thinking, Xiao Ming’s mother knew that Xiao Ming’s test scores were not good, and through subdivision thinking, she also knew that he failed the English test, but she still didn’t know why he failed the test. By talking to Xiaoming and learning about the details of his exam at that time, he found that he was uncomfortable at the time and couldn't concentrate on answering the questions, which caused many questions that he would have done wrong. After the heart-to-heart talk, Xiao Ming's mother expressed understanding to him, and since then became more concerned about Xiao Ming's physical condition, the relationship between them has deepened, and Xiao Ming's grades have become better and better.
If you continue to use traceability thinking to analyze, you can gradually deepen your understanding of data sensitivity and business.
4. Related thinking
In the era of big data, the core is related thinking, which is based on related analysis.
The story of beer and diapers is a classic case of correlation analysis. This story originated in a Wal-Mart supermarket in the United States in the 1990s. At that time, Wal-Mart had the world’s largest data warehouse system. In order to accurately understand the buying habits of customers in its stores, Wal-Mart analyzed the shopping behavior of its customers. Know what products customers often buy together.
The detailed raw transaction data of its stores are collected in the Wal-Mart data warehouse. On the basis of these original transaction data, Wal-Mart uses data mining methods to analyze and mine these data. An unexpected finding is that the most purchased product with diapers is actually beer.
After a lot of actual investigation and analysis, it revealed a behavioral pattern of an American behind "diapers and beer": In the United States, some young fathers often go to the supermarket to buy baby diapers after work, and 30 of them %～40% of people also buy some beer for themselves. The reason for this phenomenon is that American wives often tell their husbands to buy diapers for their children after get off work, and their husbands bring back their favorite beer after buying diapers.
In most cases, once we have completed the relevant analysis and are no longer satisfied with just knowing the "what", we will continue to go to a deeper level to study causality and find out the "why" behind it .
5. Hypothetical thinking
When we do not have enough data and evidence to prove something, we can make a bold assumption, and then carefully verify it to verify whether the assumption is true.
For example, one day, Xiao Ming went to buy fruit and had a conversation with the aunt who bought the fruit.
Xiao Ming: "Auntie, are your oranges sweet?"
Auntie: "Sweet, try it if you don't believe me."
Xiao Ming: "Okay, then I will try one."
Xiao Ming peeled off an orange and took a bite: "Well, yes, it's really sweet. Weigh me two pounds."
This story is just a simple analogy, and there is no need to go into details. From this, we can see the basic thinking process of hypothesis testing. First, Xiao Ming puts forward the hypothesis: oranges are sweet; second, randomly select a sample; then, test whether it is really sweet; finally, make a judgment to confirm that oranges are really sweet, so Bought.
In data analysis, the technical term of hypothetical thinking is called hypothesis testing, which generally includes four steps, namely: proposing hypotheses, drawing samples, testing hypotheses, and making judgments. We will not expand on those technical terms here.
6. Reverse thinking
Sometimes, we need to break the conventional thinking mode and think about the problem from the opposite direction. We continue to tell Xiao Ming's story.
Once, Xiao Ming went to buy tomatoes, and there was another conversation with his aunt.
Xiao Ming: "Auntie, how much are your tomatoes per catty?"
Auntie: "Two dollars and five dollars."
Xiao Ming picked 3 of them and put them on the weighing pan: "Auntie, help me weigh them."
Auntie: "A catty and a half, 3 yuan and 7 gross."
Xiao Ming removed the largest tomato: "You don't need so much soup to make soup."
Vendor: "One catty two taels, 3 yuan."
Xiao Ming picked up the largest tomato that was just removed, paid 70 cents, turned around and left...
You see, the use of reverse thinking can sometimes have unexpected results.
7. Deductive thinking
The direction of deductive thinking is from general to individual, that is, the premise of deduction is general abstract knowledge, and the conclusion is individual specific knowledge. The main form of deduction is "syllogism", which consists of three parts: major premise, minor premise, and conclusion.
Take a common sense in physics as an example.
Major premise: metal can conduct electricity.
Small premise: Copper is a metal.
Conclusion: Copper can conduct electricity.
It can be seen from this example that the major premise is the known general principle (metal can conduct electricity), the minor premise is the special occasion under study (copper is a metal), and the conclusion is that the special occasion is classified under the general principle. Knowledge (copper can conduct electricity).
8. Inductive thinking
The direction of inductive thinking is the opposite of deduction, and the process of induction is from individual to general.
Still take metal as an example.
Premise: Gold can conduct electricity, silver can conduct electricity, copper can conduct electricity, iron can conduct electricity,...
Conclusion: Metal can conduct electricity.
The process of data analysis often involves first contacting individual things, and then summarizing, generalizing, and then deductive reasoning, from general to individual, and so on, continuously accumulating experience.
to sum up
This article summarizes the eight kinds of thinking of data analysis, namely comparison, segmentation, traceability, correlation, hypothesis, reversal, deduction, induction, make full use of these thinking, whether it is work or life, I believe that it can create more value.
