Authentic Big Data thinking right here!

According to Alibaba vice President, data Committee president Cheping, this is the normal social data, it determines the era of data analysis of the characteristics of the tide-Leng noble grounding gas, unrestrained flavor depth.

To know, the Windows95 era of data analysis, MB magnitude; Google search era of data analysis, GB level; and in the social era, data analysis is TB, often thousands of times the number of growth, to now, nature has become a "big" data analysis, large in the number of data levels, but also large in the various dimensions of data, in addition, more emphasis on veracity (authenticity), Velocity (in time), in turn, large data analysis can really "be implemented", "can be explained" and "data accurate, stable."

A lot of ideas? Sorry, little Shang can only tell Shanyu, these concepts are the basis of large data thinking, without them, talking about what big data are roots, source. This is the second time to listen to Shang teacher sermon, in Shanghai, "Forbes" forum. Unlike last year's interview, this time a number of industry veterans of the discussion, the collision of ideas, showing large data thinking in practice sparks.

The question is the answer, you understand?

The question is the answer? Yes, big data thinking is about asking.

Taste said a case--before, the southern grid has been consulted, they collect the time of each household electricity data useful? Product feel considered, ask each other, can use different periods of electricity consumption data, to estimate whether the large household appliances failure, the grid people say yes, so this can be derived from the service, Remind homeowners "electrical appliances may fail", early repair, to avoid the occurrence of short circuit, fire and other hidden dangers. and further, if you can understand the brand of household appliances, models, more can tell businesses, the geographical distribution of related products, the actual failure rate and so on data, these to their production, sales and has great significance ... Isn't that big data thinking?

The founder of the network of points, Guo Aerospace, shares the first case, he has been to the famous Peer-to-peer network loan peer Lending Club to study, in the early years, the Lending Club has been collecting some seemingly useless data: for example, the time users surf the Internet, fill in the loan application form, but later, the time is long, The Lending Club found that these statistics are closely related to borrowers ' default rates--people who are always on the internet in the middle of the night looking for loans, they may not have very stable jobs, and those who fill out the loan application form soon may be less cautious in their lending, and their motives are worth examining ... Seemingly useless data can finally show great energy.

So, to see the data is not important, the key is how they should be used to show the value of the place, so that the boss more understanding of the business operation of the various sections, so that the wind control more risk-controlling angle, so that runners know more about their body parts of the State ... Well, the big data is for you.

Between the hypothesis and the answer

Two more jokes:

If Jing Dong founder Liu to you to "weekly", what kind of weekly report do you want to make? Tell him: "The tomato as the representative of Raw fresh do very well", or: "Tea business promising"? It's not right, you should take weeks as the unit, looking back on the tasks that Liu arranged before, where to do well, why good, where there are problems, Why is there such a problem, what are the solutions, what are the best solutions, why, what are the important changes in the current electric quotient environment, and what kind of adaptation changes will be needed? Without accurate assumptions of large data analysis, even with the results, can not convince Liu such a big guy, because the first direction is wrong, assuming horseshoes, the results striking. Use big data to curry favor with the leader? Beware of kissing the horse's hooves.

Another internet company executive said that Ali investment Youku, the next day the stock must fall, "This is my big data." "Yes, the next day the cool stock fell, but this is the big data?" true nonsense, do not rise on the fall, that is, 50% probability must guess right. Large data analysis is not a guess right, to the second "guess" is really big data analysis, more core, to "guess"--rose, how many points, fall, how many to fall. Otherwise, than the Shibi count is not reliable, how can it be large data? Small Shang personally think, pretend to understand, it is this group of people dirty Big data of the beautiful vision, is "bad currency expel good currency" variant, SB XI XI. (Please forgive the little Shang and curse again)

Visible, do not understand the hypothesis can not find the answer, the answer can not prove that the big data thinking, in a rigorous mathematical system, bold assumption, careful verification, adjustment, alignment, is to seek large data positive solution, although it is doomed will not be flat.