Hequan: Large data value comparable to oil

At present, the developed countries have begun the strategic deployment of large data, but most of the large data in the country has not been effective value mining, is in the initial stage. The correlation analysis shows that China's large data market will grow from $77.6 million trillion in 2011 to $617 million in 2016 and a composite growth rate of 51.4% over the next 5 years. Experts said that the use of large data mining to enhance the government management function and enterprise decision-making capacity, innovative development model will have far-reaching impact, enterprises should speed up the strategic deployment of large data, to seize the opportunity for large data applications.

In the recent Guangdong Internet Conference, academician Hequan of Chinese Academy of Engineering has invoked a large number of scientific data and cases to show that large data is the means of production, and that the industrial change caused by large data has begun.

-Win weekly reporter Zhu Weiwei

Big Data is new wealth

Win Weekly: Now all walks of life are talking about big data, how should we understand large data?

Hequan: Large data refers to a collection of data that cannot be crawled, managed, and processed with conventional software tools within the allowable time. The standard for large data scales is constantly changing, and the size of a single dataset is currently in between data TB and data PB. For example, security monitoring data volume, now video surveillance cameras are widely used in several cities in China, the main roads, hot spots, subways and residential areas of security surveillance, in Beijing, there are almost 800,000, in Beijing as long as a person went to public places, at least 8 times a day.

Enterprise data is also many, the world's enterprise storage data total reached 2.2ZB, the next year 67%. 10KB is about the equivalent of a single sheet of paper filled with text, and in the case of a stack of paper, 100 of TB is stacked equivalent to the height of the Empire State Building. Taobao in 2010 has 370 million members, online merchandise 880 million, daily transactions more than tens of millions of, its one-day data generation more than 50TB, storage 40PB, only (this year) November 11 to accept 215 million users shopping, processing transactions 105.8 million, peak up to 90,000 pen/minute.

Win Weekly: What is the application value of big data?

Hequan: The application of large data is very extensive. IBM Japan's economic indicators forecasting system, which searches the Internet news for 480 economic data affecting the manufacturing sector, calculates the adoption of a manager's index (PMI) estimate. Using Google's mood analysis tool, Indiana University has dug out the user's mood for 9.7 million messages posted in 2008 3-December for 2.7 million of users.

This is a business involving management, development, sales, support, service sector applications, the past, the association between these departments is not enough, now through large data analysis, can be linked to a number of departments, can greatly improve the efficiency of enterprise management 40%-60%, Traditional enterprises have a data warehouse is very good, but only the data warehouse is not enough, need to mining content, extraction services.

Big data is also important in agricultural applications, with a climate company in Silicon Valley that obtains decades of weather data from databases such as the US Meteorological Office, and makes precise icons of rainfall, temperature and soil conditions and crop yields over the years, thus predicting next year's output for any farm and selling individual insurance to farmers, Weather companies will pay in time if bad weather fails to predict damage to crops.

What kind of customer buys most things? Women。 Which customer group is a gold customer? Pregnant。 A company to find some characteristics, pregnant women generally buy some no irritant cosmetics, as well as calcium supplements and other health care products, according to these (CAN) judge a woman is not pregnant. They put some pregnant product ads in the ordinary advertising consignments sent to customers, to the customer segmentation group sales.

A classic case of large data is that Wal-Mart through the consumer shopping behavior and other unstructured data analysis, to understand customer shopping habits found that when the young father generally buy urine is not wet, usually to reward themselves to buy a beer, so put the two products together to sell the effect is very good.

The Wall Street (DE) Winter Capital markets Company is analyzing rumors about 340 million of global Weibo accounts, judging people's feelings. People buy stocks when they're happy, and they sell stocks when they're anxious. To decide whether to buy or sell shares in the company, the company received a 7% per cent yield in the first quarter. Recently, many of our small and medium-sized enterprises from the bank can not borrow money, because they do not guarantee, Ali company based on the changing of Taobao screening of financial health and integrity of enterprises, so do not need to guarantee loans, has lent more than 30 billion yuan, bad debt rate of only 0.3%, can be seen Taobao screening is very accurate.

Operators have a large number of mobile phone data, through the mining of mobile phone data, not for personal face is cover and group behavior, can be analyzed, real-time dynamic mobile population source and distribution, travel and real-time traffic flow information and congestion, items selling situation and so on.

Obama has just successfully campaigned, and his team collects voters ' interests and concerns through social networks and microblogs, using software to analyze and build voter profiles. For example, most postings on Facebook or Twitter by a voter are about environmental and medical costs, and they can send a message from Obama about environmental issues to give the electorate reason to support the president's re-election and get fresh feedback from voters.

As a result, the World Economic Forum reported last year that big data is new wealth, worth as much as oil.

Win Weekly: What are the special requirements of mass data in management and analysis processing?

Hequan: Large data management, from different places, different standards, large data, structural forms, real-time and other requirements, increase collection, and integration difficulties. Data storage, traditional centralized database, data Warehouse system can not effectively deal with large data storage and analysis, the need for distributed processing. Generally speaking, the first data acquisition, from the mobile internet, things, internet access to data, and then sent to the storage system, the light can not save, if there is no label, the future can not find, so the need for tagging, summary, convenient quick inquiries, and finally effective statistical analysis.

We are now using low-cost distribution processors, low cost How to ensure reliability? It is often stored in a cloud, and a data is divided into several different places. We start with different types of data to classify, different business data classification, different colors to represent different types of business, to different places, only in order to further optimize the data processing. The challenge now is to decompose the data, and the structured processing needs to be correlated.

Big Data triggers industrial change

Win Weekly: The conclusions based on large data analysis have influenced the behavior of many individuals and enterprises, specifically, what kind of industry changes have been caused by large data?

Hequan: Industry changes triggered by big data have begun, such as the drive for intelligent production and the revolution in wireless networks. Gartner's 2012-2016-year it trends show that large data, cloud storage, mobile applications, social networks, and network security will affect the IT industry over the years. Now that social networks are an important source of big data, Gartner points out that big data makes the content analysis of social networks increase by 45% a year. As an important gateway for large data applications, intelligent terminals can produce data that is closely related to consumers, and the global billions of users are enough to make this data invaluable. Mobile internet based on intelligent terminals has become one of the commanding heights of large data market, and mobile internet and large data have promoted the transformation of it enterprises. Now the terminal PC is thinner, the handset intelligent terminal is bigger and larger, the gap between the two is getting smaller. One hand is a smartphone, and two of them are tablets. The properties of communication in intelligent terminals have been declining, and the properties of consumer terminals have been gradually increasing.

We can take a look at the traditional mobile phone companies, Siemens, NEC are now out of the mobile phone market, Apple, Samsung, Google has never been a mobile phone companies actually came in. (and) domestic enterprises, Ali do mobile phones, millet to do mobile phones, Baidu, Shanda, Renren also intends to release mobile phones, not to say that the mobile phone hardware can make money, the purpose is to let his mobile phone users on their web site, follow-up through the flow from the ads to cash.

Apple captures the opportunities of mobile Internet, combines products and services to achieve industrial chain integration, not only technological innovation, but also business model innovation, Apple-built platform to control mobile data. Nokia's ability to meet its performance on functional handsets and to slow its transition to smartphones has been a slogan for the transition to the mobile Internet in 2006, but a lack of action. Sony Ericsson missed the combination of music platforms and data, had been dumped by the market, and the company bought all its shares and Ericsson withdrew from the handset market. Ericsson is now the world's fifth-largest software company, not the largest hardware company, and its share of professional services is 38%. Microsoft is also high-profile in communications and large data areas, 8.5 billion of the acquisition of Skype, social networking, internet telephony and video conferencing, the acquisition of AOL's patents, the layout of mobile search, the release of tablet computers, into the hardware sector.

Google in 2011 to buy Motorola Mobile 12.5 billion dollars, into the mobile communications sector, integration services, the social strategy to the first place. Google also developed large Google data, can allow users to upload their data free of charge, Google to help you analyze the same time also mastered your data. Facebook, the biggest corporate social network, is concerned about the occupation of social networks, so it buys mobile-phone location providers and will launch its own-brand smartphone. The two owners of the acquired company were young, the company was bought by the photo agency, a total of 13 people, set up 1.5, sold to FACEBOOK10 billion dollars, on average to sell more than 80 million dollars per person. Facebook believes that photo-exchange is an important part of future social networking.

Win Weekly: What is the key link in the value chain in this industry change?

Hequan: Large data age, in fact, the most not to sell hardware to make money, the most money to do the service, now the entire IT industry value chain to the content supply, to the data provider transformation, the entire industry chain to this direction transfer.

In the final analysis, data is the means of production in the present. International attention to large data, mobile Internet, the next generation of Internet, IoT, social network, cloud computing, large data combination will set off a new wave of network business development, the value of large data sharing and utilization will create huge wealth, will bring new industrial development.