That's how I use the data: two Doctors

Today begins with a little story. Said there were two doctors, all of them engaged in all kinds of surgeries, and we took these surgeries and sorted out some surgical statistics, as follows (pure fiction)

Doctor A, male surgery 16 times, failure 3 times, failure rate 18.75%, female surgery 1000 times, failure 39 times, failure rate of 3.9%

Doctor B, male surgery 10 times, failure 2 times, failure rate 20%, female surgery 1200 times, failure 47 times, failure rate of 3.92%

Seeing this data, let's evaluate the skill of two doctors, it's easy for us to get: male surgery, a doctor's 18.75% is less than Doctor B's 20%, and women's surgery, a doctor's 3.9% is less than Doctor B's 3.92%, so, whether in that respect, a failure rate is lower, is a better doctor!

But the story is not so simple, if we treat all the surgery as a whole, then the overall failure rate of Dr. A is 4.1339%, and Doctor B is 4.0496%, strange things appear: Although Doctor B, both men and women lag behind with a doctor, but the overall is more excellent!

I did not play the number game, did not deliberately calculate the wrong proportion, you can personally recalculate, a indeed all aspects are excellent, but the overall is behind!

No doubt, this is a problem! But this is not just the strange question itself, we advocate to look at the negative side of the problem. How do people evaluate things? I think the vast majority of people from the traditional direction of this article to draw a better conclusion, which essentially reveals a good plan: the strength of the weaker a, you can follow the same article, to overcome the strength of a stronger than their own B!

Haha, this is like Tian Bogey horse racing! Everyone says that the mall is the battlefield, but not everyone is going to think how to fight every battle of life! Let us not dwell on the conclusion of the horse race, let us continue.

Take a closer look at how a is done, we can see that the total number of male surgery is significantly less than women, this may be an objective fact, and the failure rate of male surgery as a whole far more than female surgery! This suggests that male surgery is the so-called business difficulty. What do you do with a B? A significantly enhanced male surgery research, making the failure rate relative to B down 1.25%! This is a key move to make a look at the relatively small frequency of male surgery, usually we call this--details!

Well, the conclusion is, we're from the digital point of view: attention to detail can defeat the strength of more powerful opponents! This can be well mapped to our software development, and usually we have two choices

1, more general development of more functions!

2, more detailed development of the details of the function!

Assuming that each person's ability in a short period of time is not to improve the quality, then the relatively constant ability in the attention to the details of the choice of the attention of the leadership, be recognized, and thus get promotion opportunities! What's more, people who develop more features will get more work and complain about their results, maybe your ability is good, but because you choose to ignore the details, your overall score will be very low. I don't know, do you understand that?

Well, the story is supposed to be over, as usual, I'm talking about the idea of the data. Generally speaking, everyone is China, is relatively blind, and only through the calculation of data, will be rational understanding! Previously commented on Sina Weibo, he said that the role and responsibilities of the data analyst are all over again: "How to show the data, how to get a chart"! I would like to say: "Do unto others others", the data analyst himself did not think from the point of view of the data, how can they use the data to fool others? My understanding is: The good data is that it can be used in any unexpected place! At the right time, in the right situation, it always appears in our confused place, guiding us where to go, people put this at the key crossroads, access to data methods called: Science!

Come on, the details decide success or failure!

See more highlights of this column: http://www.bianceng.cnhttp://www.bianceng.cn/Programming/project/