Traditional media want to borrow large data against the wind again as the scientific application of prevention mistakes

For the traditional media, if we want to realize the digital transformation, we must establish the Intelligent Information Service platform based on the large data technology. So what is the big data? What are the limitations of traditional media entering big data? And what can be broken? In this regard, the State Administration Institute of Social and Cultural Research Department of senior economist, management Doctor Guo Kanzhong did an analysis.

2013, is a big data year, big data is bringing revolutionary changes to traditional media, although traditional media also take various ways to actively deal with the challenges of large data, but because of the lack of awareness of large data, leading to practical deformation, misunderstanding.

Traditional media large data practice benefits

First of all, through large data analysis in time to capture the traditional media users of the relevant reading information, improve the user experience. At present, both the media and the traditional media, the user experience is its short board, and large data analysis can be analyzed by the user's concerns, needs and other data to better meet the needs of users, and thus enhance their user experience

Secondly, actively carry out the experiment of news data. At present, the user is more interested and concerned about the visualized data, some important reports of traditional media pay more and more attention to the visualization of information.

Third, the use of large data technology to develop public opinion management related products. At present, our country is in the social transformation period, the public opinion management demand is very big, traditional media because has the strong news excavation ability and the dissemination ability, may use the Big data analysis method to exert the public opinion management business. People's Daily People's People's Office in the public opinion management has done very well, the annual operating income of billion.

The misunderstanding of large data practice in traditional media

First of all, still uphold the "content is king" concept. Large data era, simple content has been difficult to form a commercial closed-loop, only the use of large data technology to achieve information and user personalization, customization needs to achieve business closed loop. However, from the traditional media practice, the mainstream concept is still "content is king", leading to the Internet concept and technology is not enough attention, large data platform is difficult to achieve.

Secondly, digitization is mistaken for data. At present, many traditional media in practice is only through electronic version, Internet Web site to achieve content online rendering. The essence of large data is to establish the internal relationship between different data and establish the connection between users and information, through data mining and analysis, find out the correlation between different things, and then realize business value. Because the traditional media practice only completes the material construction part, is still far from the real data.

Third, the mistake of the news visualization as a data. At present, many traditional media in the news, often with the help of visual tools, but most of the visual news only to pursue the beauty of the news, and not fully reflect the nature of the logical relationship between the data, can not effectively promote the user to think, but also can not effectively present its purpose.

Scientific understanding of large data

First of all, large data refers to the massive, high growth rate and diversified information assets that need new data processing mode to collect, store, manage and analyze their contents. Large data is a collection of new ideas, ways of working, and tools, not just tools.

Secondly, large data has the characteristics of online, mass, totality, unstructured, real-time. In linear, that is, large data is always online, can be called at any time; massive, that is, large data scale, the current usually refers to 10TB the amount of data above the scale, the whole, that is, large data to take the whole thinking, rather than sample thinking; unstructured, that is, the wide variety of large data, including not only traditional relational data, and includes raw, semi-structured and unstructured data in the form of web pages, video, audio, e-mail, documents, and real-time, that is, large data can react in real time. For example, enter a keyword in the Google search box to instantly render.

Third, large data represent new ideas and thinking. Large data can deal with "causation" and deal with "correlation", that is, not only can answer "why" and can answer "what". In the small data age, only by sampling the way to answer "why", and large data can be a full sample of the way to answer "what", that is, to find the relevant relationship.

Four, the key to large data is intelligent, that is, the use of effective tools for data mining and professional processing, and then through the "processing" to achieve data "value-added", and thus achieve profitability. At present, the methods used are mainly data mining and comparative analysis, the main related technology is mapreduce and Hadoop as the representative of the Non relational data analysis technology.

The implementation of large data depends on the availability of the data, the science of the model, and the refinement of the view. At present, many data are difficult to obtain in China, which makes it difficult to realize real large data mining and analysis. Secondly, the scientific model directly determines the quality of data analysis, which requires a superb level of modeling; Thirdly, the original and high quality view based on data mining, which provides the basis for decision-making, is highly dependent on high quality data interpretation, This reflects the value of industry experts.

VI, large data and cloud computing organic deep integration. Cloud computing has already implemented data Analysis Services, both of which are accompanied. In addition, when future web-based semantic networks replace web-based Web sites, large data will become the mainstream of our access to information.