That's how I use the data: 1 Baskets of eggs

A basket of eggs in front of you, so you pick a few to take back to eat, what will you do? I think the general will be like this: can not say all take away, this is not good, but also to see which several large, and then pick those few! Today want to share with you a little bit, by the selection of the Data driven meaning!

Say Han, maybe "data-driven" is not my proposition, but these four words are really appropriate, very consistent with my thinking, ready for long-term use, however, the problem is really difficult to explain what it is, tried many angles to explain it, have not made it clear, of course, this gives me time and time again to reflect on the opportunity, It's not bad! Today this topic is to continue to take the example of eggs to talk about data-driven, may be your dish, may not be your food, but I decided on countless dishes, you will find the one you want! Ha ha! This is in itself a data-driven approach!

Let's get back to the egg topic! Review the egg owner's statement: "Pick a few to take back to eat"! Consider two questions:

1, he said pick a few, but did not say "pick three" or "pick five", what makes you finally take 3 or 5?

2, he said pick a few, but did not say pick which several, or so asked "pick what meaning"? But this does not affect you to choose the behavior and rules of the egg, this rule how to come?

Well, the question is finished, some people may want to say that these two questions too simple, can be summed up to two words "common sense", because common sense others let you pick a few, then generally you also take 35! According to common sense, pick is to choose what you like, if you like the big to choose the big, if you like small, and no one to stop you!

Oh, it is "common sense", this is so simple, but also stifle the imagination of the thing! Let's muddy the water and see what "common sense" should do? For example this person altogether has 3 eggs, lets you pick a few, you choose according to the common sense 35, then three all took away!

You might argue that if this person has 3 eggs in total, according to "common sense" you should just take one! Although it seems that the common sense is strong, but also reflects that it is not an oil-saving lamp, with considerable complexity, it is not easy to master it, this is a "behavioral art", may take it for granted that it is common sense, but it is already complex to not common sense so simple!

What is this? Let's turn to think, in fact, common sense is not let you take 35, nor let you take one, but let you take a very small percentage, that is, do not take too much! With this in mind, the question is converted to: Draw up a suitable percentage and take care of this common sense!

Let us talk about the second question just now! We say pick a bigger. So, which one is bigger? Perhaps you want to despise me, most can not see Ah, is a bigger ah! Well, it's a feeling thing, the simpler, the more accustomed, the more stifling the imagination, if we let the program execute the process, do you write: pick (larger)?

Well, when it comes to software development! In fact, "big" is not difficult, right? According to the "data-driven" understanding, we have already mastered all the data, to find you the largest, not on the line? Of course, in addition to looking for the biggest, there is a more interesting way to share with you:

1, master all the data, calculate the average U

2, set a suitable x, whenever an egg weight greater than u+x it is "big"!

This method simulates the human vision, calculates the average value u, is the basket egg most long what kind, then is bigger than the average X unit's egg, is bigger than the average egg to arrive certain degree of kind! Hey, is not more than the "big point" to say more "reliable" Ha!

Of course, this method is not perfect, the evil of mathematics gives us a standard set X method:

1, master all the data, so that the weight of each egg minus the average u, get a value, and then squared this value, get D (i)

2, the average of all d (i), get a value, the value of the square root to get X (in the formal statistics tutorial This amount is "sigma")

Let's see what this is doing! Each weight minus the average is used to mark how big the difference between the eggs and the general size is, and the square is so large or small that it cannot be offset, we all understand that the average is to make these differences average, the last open root is because the square before, here the unit again converted back!

In fact, I can tell you this is to ask for "standard deviation", but why to describe the process of the solution in detail? Because, although this is the recommended approach, but "data-driven" is so attractive, it is not a "statistical" drive, the imperfect rules still work, you do not have to master the evil of mathematics, because it can not only drive the business, but also to drive the "data-driven" itself! We can try the appropriate simplification rules as long as you like! For example, here, I do not square, but to seek absolute value, so there is no need to open the root! Although there are some flaws, but the overall direction, it has long been able to drive software to use data to improve the rules, the other, who cares?

Well, get x, theoretically more than an X is very big, more than 2 x is super large, greater than 3 x, that may not be eggs, but a goose egg accidentally mixed in!

Let's recall that this basket of eggs is just themselves and can be used to illustrate how to take a few as "common sense". You can use the weight of each egg naturally determines which is larger, or even a goose egg in the inside, do not tell me, look at the weight can be judged, I would never think it is a super big eggs! This is called exception data in software, I mean, do you understand?

Well, the story is coming to an end, according to the convention we talk about data ideas! One of the most important ways this article series emphasizes is to interpret business with data rather than artificial designations, such as the eggs here, the traditional software development must ask more than how many grams is big, you let the customer how to answer you? And as new eggs join in, big standards change at all times! As the eggs are taken away, the big standards are changing! Let's just say that this standard is getting too fast, but it always seems to fit the need! This is "Data driven"! It's cool, you try?

Come on, write the program with Data driven, blue bottle, good absorption!

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