Application of large data in O2O

April 24, 2014, Baidu held the fourth session of Technology Open Day, the main topic of this meeting is "Big data engine driving the future", this is the first September 2012 "cloud Computing", July 2013 second "box Computing" and November 2013 to "Light Application", "API Open" As the main content of the third technical opening day, after the Internet and traditional industries in the context of the deep integration of the era, held another technical meeting, the core of the conference discussion fell on the "big Data."

Baidu chairman and CEO Robin Li once again reiterated that "the Internet is accelerating the elimination of traditional industries," and said that "from last May to the present, the Internet is almost the first to subvert the media, and later subversion of retail, and then tourism, and then finance, the Internet and traditional industries to become a trend." and "Technology will change the Internet", "the quantitative changes in technology development will bring about the transformation of the world"; how technology (especially large data) plays a role in the O2O tide is a question that people ponder, this paper puts forward some thoughts on the application of large data technology in O2O field, and looks forward to discussing with

Basic information and characteristics of large data

The concept of large data was first put forward by an article published by the International Top journal Nature, "Big data", and it was predicted that the information Marvell era would soon usher in the DT (data Marvell) era. Large data has the following characteristics:

As the parallel computing power increases and the cost of data storage decreases, large data is measured in PB or Eb (1EB equals 1024PB,1PB equals 1024TB) and is still growing at a rate equivalent to the "Hardware Moore's Law" (doubling every 18 months).

Large data have 4V characteristics, namely large scale (Volume), rapid Change (Velocity), multiple species (produced) and low value density.

The research of large data computation should focus on 3 I: Approximation (inexact), Increment (Increment) and inductive (induct).

How large data technology can play a role in the O2O field

Deep analysis and dynamic mining of large data, because the data sample is large enough, will form a large number of things to reflect the nature and appearance of the law, these laws will be "complex disk-type" reaction to the process of data generation, and for the traditional industry to provide unprecedented depth optimization and intelligent decision-making, until the formation of operating methods and products of the "subversive"

Take the big data engine of Baidu for example, the technology contains 3 important components. "Open Cloud", "Data Factory" and "Baidu Brain". Using a deep neural network to build the "Baidu Brain", its computational capacity is now equivalent to 2-3-year-old children's IQ.

Shongju, director of the Ministry of Transport and Science and technology, referred to the "smart traffic in the big data age" and threw in a lot of new ideas about traffic, data traffic, handheld traffic and human traffic. For the Traffic management department, real landing to do things far more than the concept of "human traffic", such as the small part of the meeting in the emergence of an idea, whether to allow each bus driver's mobile phone to share their GPS location, the location of the data published to a bus location query platform, each public transport by the people mobile phone installation " Bus location Query App ", so that you can subscribe to the way you want to see the bus when it arrived, reasonable arrangements for waiting time, so that the office workers do not have to stand in the yet bus station because of waiting for the bus and shivering. Of course, if in the "Bus location query App" Implanted ads, I think the effect should be more than the bus in the LED bus media is much stronger.

What should be paid attention to when the enterprise uses the Big Data technical guidance O2O

In the field of large data research, the threshold is more and more high, the need for collaborative computing to achieve the best results, in the O2O area, all mobile Internet applications are not in the production of data, and data stored in each company or application of the database server, in large data background, It's not going to work. Depth of data analysis and mining, with peers, and upstream and downstream of the effective large-scale data collaboration analysis will bring great value to the entire industry chain.

Nowadays, the data management information system can not meet the requirements of the traditional enterprises, and the traditional enterprises need intelligent optimization and assistant decision based on large data. In terms of data resources, traditional enterprises should make full use of the existing large data resources, such as Tencent's social large data, Alibaba's commodity and trading data, Baidu's "General" based on the search of large data.

In the large data acquisition, can not only stay in the PC Internet, should be widely access to data from multi-channel, and even create new ways of data acquisition. such as wearable devices based on a variety of sensors, self-built regional WiFi network, LBS technology and so on.

The traditional manufacturing industry should give full attention to the optimization and guidance of large data. On the basis of the production management informationization and Process Control automation, the manufacturing industry will enter into the era of technology manufacturing intellectualization, and the traditional manufacturing industries will usher in the revolutionary "subversion" of the bloody revolution under the impetus of large data technology and O2O technology.

Through large data analysis, traditional enterprises should pay more attention to the user experience, large data will help some traditional industries to completely overcome information imbalance, data opacity, inefficient management and even trade latent rule, and believe that with the support of large data technology, O2O will move forward quickly.

Thinking about the function of Baidu Big data technology in O2O

(1) Large data storage and large data presentation mode

Baidu has built a super large-scale storage, distributed computing capacity of the open cloud, at present, to solve the domestic large data needs should be sufficient, how to protect the enterprise data assets in the premise of the promotion of enterprises in the industry joint collaboration for large data mining is a problem to consider, it is recommended in the implementation mechanism, The result sharing rule aspect carries on the plan, thus uses the technology to promote the industry most enterprise common progress.

(2) Building and sharing large data optimization common technology

At present, Baidu Data Factory has been in the transport, medical and financial fields with industry experts to cooperate, but it is not possible to solve all the problems of the industry, so Baidu should open up the cloud and data factory on the basis of further open up large data mining and analysis of common technology, the common technology modular, the formation of To O2O entrepreneurs and traditional industry experts, O2O entrepreneurs and traditional industry experts to use these large data analysis and optimization tools to form a series of large data based intelligent optimization and decision-making program to help O2O rapid development.

Baidu has emerged from the Chinese search engine market. For the large data technology and O2O technology optimization and promotion, Baidu is the advantage, but also the responsibility.

The author of this article, WB expert author, Micro-BO: WB-billion European network; reprint please specify the author's name and "Source: billion European Network"; The article content is the author's personal opinion, does not represent the billion European network to agree or support the opinion.

Statement: billion European network to respect the industry norms, each article is marked with a clear author and source; billion OU net original article, please reprint must indicate the article author and "Source: billion European network", do not respect the original behavior will be billion European network of responsibility.