Facebook's data model

In the opinion carnival around Facebook listing off, I noticed a funny phenomenon.Facebook has been the Internet community as SNS concept, listed in the eyes of investors, has become a data concept.

CITIC Securities advised investors to "listed on Facebook, to lead the Internet into the era of big data," by Facebook's potential, not cast everyone net SNS concept, but cast the TRS (unstructured information processing) this concept of large data .This is very typical "Facebook species" on this issue, there is a layman and expert understanding of differences and a completely difference.

So, what exactly is who in the watch, who is subtle? Unfortunately: the Internet community to watch, and investors but subtle.

Facebook First, large dataI used to have a view, or call the different views, and Internet peer exchange was very strenuous. They always think that, Facebook is the SNS; I do not think, Facebook data. I said: "The domestic industry is always mask as a the SNS relish. In fact, this is an amateur to watch the view. Facebook is really the SNS, but his real core competitiveness, the data on the core business."

Heterogeneous view of this, the project manager, the advertised SNS concept of project manager, received little resonance. This time, investors finally encountered a bosom friend. I think that person's point of view of capital, than the project manager's point of view, is closer to Facebook's reality.

First of all, the capital is broader than the project manager perspective, longer in the Facebook era background seen, rather than the project manager that purely business point of view.Capital to Facebook into the two larger context: one in the era of judgment to recognize the value of social networks (SNS), mining lead the Internet into the era of big data, and to promote the development of "big data" industry. 2, in the judgment of the industrial chain to recognize the social network Facebook was the first to enter the era of large data, will further lead to other Internet areas of data applications, the excavation of the user value will drive the development of the industrial chain of "big data".

The project manager noted that the Facebook SNS SNS in the greater scope of what to do with, not the SNS as the goal itself.

Second, the capital longer than the project manager through the SNS phenomenon, see the nature of the data.

Capital people to Facebook this "special" on the "General": 1, capital people have realized that the SNS is just data in a special case of the acquisition side: large data refers to the "huge amounts of data + complex data types, while the socialnetwork (SNS) is precisely every second to generate vast amounts of unstructured data (text, applications, location information, pictures, music, video, etc.), is a typical "big data" system; the SNS only data The core of an application: "data" is data mining and multi-faceted value. Social networking (SNS) the value of mining itself is a "big data" and the important applications of business intelligence applications. 3, Facebook represents a one-on-one consumer-driven model transformation: Facebook user data contains a huge commercial value. Users comments, upload pictures, music, video, etc. are typical of unstructured data, which contains the user propensity to consume, "Data mining analysis of advertising can greatly improve the precision put in effect, is conducive to the Facebook Developer more compelling applications on the user, and user behavior to predict a number of industry trends, had enormous commercial value.

In contrast, the project manager for the understanding of Facebook, too concentrated and functional details of the transition has nothing to do, and less understanding of the meaning of the Facebook business model in transition.

Zuckerberg, even the capital people can read through his heart. Capital only at the level of value, the interpretation of what Facebook is. Zuckerberg has repeatedly stressed that Facebook was created to not become a company, in the sense level, read Facebook's mission.

Facebook will not SNS afterFacebook is far from the SNS, the SNS can barely be called an endowment of Facebook. Endowment for the SNS or even just a variety of possible endowment one. This commercial analysts see out. John Battelle of Federated Media John Battelle seen, some shift of Facebook, ongoing and past forecast. The company is trying to redefine itself, not satisfied with the narrow aspects of social networking sites , which is just outside the understanding of it.1, the operating system of people's lives

Bartley most promising a new direction, is the life of the operating system.Bartley said that all companies are rushing to become the operating system of people's lives, they want to become such a central place, the user participation and all the data stored there, and then they do anything to need to use these central.Zuckerberg said: the world's information infrastructure should be similar to the social map.

Zuckerberg's vision has crossed social, from the social to be inspired, be extended to a map similar to "the world's information infrastructure. This statement is not as good as Bartley "has become the operating system of people's lives" in place. In fact the original meaning of Zuckerberg may be closer to the mean, because people share more, they can get more information about products and services through their trust, he said:. They can more easily find the best products, and improve the quality of life and efficiency. Clearly in the emphasize the meaning of life.

The world's information infrastructure "with" life operating system "may be regarded as the prototype structure of the large data behind, comparatively speaking, the former is more focus from the object to grasp the overall structure of the data, while the latter is more focused on the main grasp the overall structure of the data . For large data, what is the life of the operating system? This refers to the reconstruction of life with meaning, the data to reconstruction of the meaning of material. Focus, the significance of the main aspects of the data will have a good or bad: the data will ultimately tend to the significance, to become wisdom; final departure from the meaning and become stupid. So, the wisdom of the earth or the wisdom of the city or not the data accumulation-oriented life seems more meaningful.

The structure of industrial society, there is no focal point on the meaning above, but focus on the value. The value and significance of the relationship, the relationship between means and ends. Valuable not necessarily meaningful, for example, money is valuable, happiness is the meaning, but the money does not necessarily happy. In other words, to master the means to achieve happiness, but not up to joy for this purpose. The infrastructure of the industrial society, human nature, and value, are fully socialized, very professional; but with significance, are in small-scale production, extremely amateur. This makes industrial society is not perfect, and can easily become a highly developed means of institutional forgetting the object and purpose of society.

Mission of large data, not from the technical means to see, but from the human perspective, is to establish a means of purpose between the focusing system of the professional, social, and thus the system so things do not deviate from its purpose,so that the conditions of industrialization in the smallholder level of human meaning systems, become a highly developed society as a whole structure.

SNS and the life of the operating system will do with it? SNS only fishing with nets, set up a focus on the significance of the lives of the operating system hit the meaning of life This is the net full of fish, is the intent of where. Imitate Facebook are nets of life data acquisition attracted and built exactly the same nets, imitate the action of Facebook's cast net, but I do not know that action is in the fishing, the results of the network is not for fish. two net for fish, and finally hit the small tail-tail fish. As everyone knows, fishing fishing nets and can also fork fish, fishing, electric fishing and other means, such as SNS data acquisition, there are many, such as LBS, O2O, payment, or even line of POS machines. If Facebook one day is not SNS, it must be found in other fishing methods that can reach more fish. The fishing metaphor here is the meaning or corporate core value; fishing means of analogy here is the endowment (usually said, What do you do, which line). Enterprises to evergreen, it is necessary to maintain the core values ​​of endowment to follow changes in the environment.


2, "reshape the architecture:Customer-centric inversion economyZuckerberg, the important principle of the "reinventing architecture: We hope that by helping people build relationships, reshaping the way of information dissemination and consumption. We believe that the world's information infrastructure and social map should be similar - it is a bottom-up peer-to-peer network, rather than the current top-down monolithic structure. In addition, let the people decide how to share what the basic principles for reshaping architecture.

This remodeling of the basic principles of architecture, in fact, guiding significance for the structure of the large data.

No principles under the guidance of the large data structure is the opposite: still follow the traditional structure of the conductivity value from producers to consumers, but the old tradition with new technology. Although this service is necessary, but not Facebook data positioning.Zuckerberg then read out the following meanings:

1) remodeling of architecture inverted economic structure means that large dataFirst, we want to build relationships by helping people reshape the information dissemination and consumption patterns ". Remodeling is inverted, so-called inverted, reversed the value generated in the direction, the original producer to consumer, and now consumers to producers. This is from top to bottom, into the bottom-heavy meaning.

SNS and search engine of this inverted structure, to generate value from producers to consumers, but in turn, from consumers to producers. Consumers first in the SNS, and search engine exposure (active "production"), consumer intention information into the exchange, so as to form abstract consumer value; second step, the processing of value-added processing for consumer information from large data, which is equivalent to the consumption capital processing, so that consumer sovereignty can be the same as the capital treatment of the same remaining.

Second, it is a bottom-up peer-to-peer network, rather than the current top-down monolithic structure ".

The traditional economy and economics, the relationship between production and consumption, an important asymmetry. "Since the production of this to the consumer this under the" order to create value, producers are primarily in the commodity and exchange links, the concrete value to change into an abstract value (exchange value); the second step, the general value through capital mechanism , value-added amplification. But consumers have no such rights and powers, one can not consumers value into an abstract value, this value is not value-added, consumption is not the process of capitalization.

Large data out of technical talk about the technology of the original phase (2012-2014), to enter the medieval stage of large data combined with economic (about 2015), people will find that the bottom-up not only involves the transmission of information, more related to the change to the value generated. Become a process of empowerment of consumers through large data. I suggest that you read the "public agitation" and "innovation enabler" two of the opposite direction to empower the book, understanding the changes in the economic life of empowerment,

The third, "Let the people decide how to share what".Mentioned here that an important concept: the "independent". The economic structure of industrialization, people lost their autonomy, the most important step is the self-alienation of labor for the labor, so the people through access to information, the first mention of autonomy in the information is to be human factors, and revert to the labor, the formation of the effect of "human + labor = independent labor".

Zuckerberg due to historical limitations, and now only vaguely felt in reshaping consumption patterns in the future to replace the small boys and girls of Facebook, you will need this information-sharing process and deepen the process for consumption capitalized. This will be the subject of the data the next phase (2018 onwards).

At that stage, people will generally think about the Haier few years ago to solve the problem: the single one "direct economic model to address the data generated by the consumer, reversed a decision of production (especially the strategic structure of the capital now the United States Institute of Management Accountants (IMA) secretly watched, and sent the name Haier strategic profit and loss account for the ZEUS (Zeus), may be the data reveals a period of capitalization strategy secret passage. Yesterday, I also specifically with the information enterprise initiated by Hu Jiansheng discuss a decisive influence on its BI, Facebook at that stage, no longer progress in his life worrying.

2) remodeling of architecture means the value and significance of the invertedThe value structure of industrial society, from the value to the significance of the first people around the means of production, then the destination on the means of correction; the value structure of the information society, from meaning to value, through the SNS, and search engine positioning significance, and then according to the meaning to do something worthwhile.

Industrial society in the past, to grasp the significance of relying on a small-scale farmers. Significance of large data should be expanded into a data structure of the system. In the significance of studies of this structure to complete the task, called the interpretation of "significance". This is a mind reader. Architecture for large data, from the perspective of the main significance, it should be a mind reader system is specialized to get rid of the human Sphinx system. Operating system through this life, so that mankind may be raised from only valuable, become not only valuable, and meaningful. Human significance and higher satisfaction.

For businesses, the reason is the same. From the significance to the value of this decision direction, Zuckerberg said: In this process, the enterprise benefits: they can make a better product - that is people-oriented, personalized products. In addition to create better products, a more open world will also encourage enterprises with clients direct and reliable interaction.Said here, people-oriented and personalized, refer to the significance of; emphasized over the value of intermediate links, to achieve "a direct and reliable interaction between producers and consumers. Zhang's words, that is, a single one.

Significance need to explain, interpretation significance cycle.


Zuckerberg as saying: it is a bottom-up peer-to-peer network, rather than the current top-down monolithic structure. In addition, let the people decide how to share what.Meaning, not the producers (the equivalent of) to give, but by consumers (the equivalent of the reader, the recipient of the product) to give. Large data system by the circulation of meaning between producers and consumers realize the value and significance of the unity, the unity of means and ends.The other hand, the large data structure there on the other hand, open up the significance of the morphological, semantic and pragmatic of the three links. The level of significance of the Sphinx, you can not write. SNS data acquisition mechanism, the formation of the significance morphological industry; next will form the semantic industry, unstructured data processing industry chain; the final formation of the language industry, LBS, payments, and other means of data mining and the specific situation of a person a person's anchor. In order to crack the personalized meaning and experience of the meaning under the language level.Speaking from the artificial intelligence point of view, Facebook's computing model has unique advantages, it is all calculated instead of Google (microblogging), the kind of man-machine computing. All calculations, quite in the dialogue, the people are each other's search engine, the formation of ecological computing power. This regard there is great potential for development.

3, exploration of large data productivity inside the engineLarge data as the engine of a new era of productivity, research productivity features for understanding the future of commercial frenzy, is a basic homework. No sense of people, technology in the era of big data, is likely to become a the bolted productivity chariot dragged the bodies.

Even know nothing about technology capital, has taken note of the typical problems of the Facebook data structure "huge amounts of data + complex data types, and unstructured data. In fact, it does not involve Hadoop, the NoSQL, data analysis and mining, data warehousing, business intelligence and open source cloud computing infrastructure such as many basic issues.

Roughly the technical process of large data, the first collector of the SNS, search engines, POS machines, the massive data collection into the data warehouse and then distributed Framework (Hadoop), non-relational data heterogeneity processing (the NoSQL), the development of one-on-one business intelligence through data analysis and mining. Due to the complexity of the problem of large data, I now some personal thoughts, but mature consideration, the first not to mislead you. Or down Facebook's practice and experience, bottom-up induction.

Facebook in large data this line is a prominent one of the protagonists. Its low-cost integration of huge amounts of data for large data line praiseworthy. Facebook's data strategy in my opinion, not yet fully finalized, it is mainly focus on the development of this one of the internal data management.

Facebook released in December 2011 Timeline is a data product. Timeline is a personal self-editing by the user timeline, simply, it is in fact a personal Sphinx solver. Asked a person, who you are, it is very difficult to accurately answer. But if a small to tell you a grown up person, and then to encounter such a problem, the response of the brain, it is a the Timeline. Than the personnel file also file. An important difference is that with the personnel files, it can control the personal information only to people who want to show. Data mining to help, in theory, a person in the selection of shoes, only to show the life history and shoes for the third-party lifestyle designer to provide advice for you one on one to select shoes with.

With the Timeline, as Zuckerberg said, "Since then, your life, all on the Internet the".Life here, life and meaning is the digitalization. Namely, the soul of this part of the survival. Soul in the whole of life, how to spend money system of the tube wallet tube limbs action system, control of a person's soul, take over the command. Timeline also may be called human Erotic system. Only the Timeline is too thin, in the future the next generation of small boys and the girls will be a better way to do it.

With Erotic system (ie, personal meaning system), collected large amounts of data to the next challenge is to crack the soul. As analysts to judge: "Facebook, before several years of efforts to figures close to 10 million immigrants established contacts and ties still expansion of the boundaries of this world, the next step is more important to consider how to make the relationship between mass data more valuable. " A few years before, in this line of large data, the equivalent of Facebook, dry mining, they were mistaken as the SNS; he actually did not agree that the next step is to switch to the the SNS mine (of course he tyrants live to do the processing of raw materials, others do not squeeze he did not inevitable to withdraw).Crack soul, in theory, called the significance of interpretation, the following must break off, the first hurdle is that hurdle, and structured data to unstructured data from the structured data is equivalent to simplify the relevant dimension after a bunch of digital , equivalent to pick out the rest of the bones in flesh and blood part, pig skeleton hanging in the markets is equivalent to this kind of thing; unstructured data, the equivalent of natural language, including extended text, applications, location information, pictures , music, video, etc., which correspond to the data of the flesh and blood. Processing of structured data is equivalent to deal with the large cavity of meat without meat, dealing with unstructured data is equivalent to deal with the flesh of the ribs. Of course, the value is much higher.

The main large data research in this direction. Specific to Facebook, it's unstructured data, mainly for usability testing, eye testing, the focus of the analysis of strategic factors, user needs, competitive products, the commercial interests of factors.

The second hurdle is harder still, from the structured data, in-depth to the potential significance of the data behind the soul. Freud worked in the history of this matter to solve analytical nonsense, ineffable, the dream. But on the scale of society as a whole, for each individual in each time history, every 200 meters in the space satellite positioning in the record, pay a water bill revenue, stored in a text for analysis to unlock the parties I am not clear about the Sphinx to him and others to distinguish between personality, and then on the NATO Air Force precision-guided one-on-one business offensive, there is now many difficult issues.

Facebook is to explore in this area, is in the active period. We can see that it is along the evolution of the order of the Face-Soul-book. Face parts, the first step, the SNS is equivalent to the data in the significance of hermeneutics, called morphological. Use SNS data mining machine, eavesdropping on users to chat, then organized into structured and unstructured data. The outstanding achievements of Facebook in this regard is to effectively reduce the cost of mining and sorting through the Hadoop open architecture.

The second step, to the outside to the inside, from the data to analyze the significance of its industry positioning service processing, namely AaaS (analysis as a service, and analytics-as-a-service). Is the "soul rebel".

The third step is the formation of the book. The ancient Chinese legend, the human in the underworld there is a book, Yang albums. Tube is the person's life. Facebook to grasp the soul of each person's secret, the records into the book, "Facebook," This step is complete, Zuckerberg has become the God in charge of human fate. In the opera "Chopping judge responsible for Big Data, officials called Zhang, due to arbitrarily rewrite the life and death book, Bao voltage. This shows that post how important it is to humans.

Facebook every day, collected 4TB user behavior data, mainly through the waterfall analysis, to track the conversion of the interactive steps / loss rate, a large number of A / B, the testing distribution, observing user behavior, usage patterns, and optimize the interface interaction and operation flow. In addition to the analysis of the waterfall, Facebook data is also used to back the sexual analysis, page optimization.

For example, Facebook has a designer to the user is about to loout of Facebook at the last minute, to restore, according to the analysis of user data, find the law of their inner thoughts, which launched the transformation of the logout page, and emotional way to move people successful write-off rate of 7%. Stopped at a critical period, Facebook's blood loss, and Facebook through the dangerous period.

Facebook's data, open to a good start, is constantly exploring innovative momentum, will continue to bring new inspiration to us. But, in general, Facebook is just a first step on this long march for the large data.

The development of large data itself needs further stereotypes. Facebook's data there are also structural problems, I think is still inadequate in-depth degree. From the face of it, the problem is demonstrated Facebook a single source of revenue, and such structural defects. Facebook also limited in their bigger data, if the can industry chain opened to outside developers like Apple, as further mobilized, the future will be more ambitious.