How to wake up the sleeping "Big Science Data"

Guest: Hu Shanqing (American Chinese statistician, American hundred people meeting Washington area co-chair)

June 8, the International Science and Technology Data Commission hosted the "Big Data and Scientific Discovery International Symposium" opened, nearly hundreds of scientists gathered to discuss how to wake up the sleeping "science Big Data." Large data has entered all levels of society, science, business, social management, and so on, all are exploring the value of large data. Statistics as a science of research data, in the era of large data, its role and status is also improving.

This March, the "National New Town Planning (2014-2020)" Issued, the 31st chapter of the plan pointed out: strengthen the statistical work of urbanization, conform to the development situation of urbanization, establish and improve the statistical monitoring index system and statistical Comprehensive evaluation Index system, standardize statistical caliber, statistical standards and statistical system methods. Speed up the development of urbanization monitoring and evaluation system, the implementation of dynamic monitoring and tracking analysis, planning mid-term assessment and special monitoring to promote the smooth implementation of the plan.

What is the significance of the current statistics in society when the big data age is already coming? Where is the challenge? What is the role of statistical application in national decision-making? To address these issues, this newspaper talks with the American Chinese statistician, Mr. Hu Shanqing, the joint president of the American hundred People's Washington area.

Breaking the key link of information statistics

Science and Technology daily: How do you view the use of data in China?

Hu Shanqing: Since 2002, China has been actively integrating and developing national longitudinal data systems, especially the establishment of definitions, codes, and standards, and it has laid a solid foundation. However, there are a number of weak key links and many challenges, includes data sharing and disclosure, timeliness and quality of information, statistical thinking and design, tools such as intelligent maps, and transparency of data collection and computing methods, consistent with the application of established standards and enhanced availability and delivery of different information, and so on. Some of these problems are global and some are unique to China.

Science and Technology daily: How to understand the role of statistics in national policy?

Hu Shanqing: Over the past century, traditional censuses and newly introduced random sampling surveys have been applied to the measurement and inference of population and economy in different countries.

But human activity is continuous and dynamic, and the census can only provide a more comprehensive speed map for a given census day or a short period. Usually when the census results are announced, they are obsolete. Nevertheless, during 20th century, both methods of statistical data support decision-making, policy development and the transmission of information throughout the world.

Censuses and random surveys of population, economy, industry and Agriculture are widely held in the United States, China and other countries. For example, the U.S. government collects data from 60,000 households a month to publish monthly reports on the country's employment situation. According to these figures, the unemployment rate in the United States is only about 0.2% sampling errors. There are 120 million families in the United States, a total of 310 million people, and many of the major inferences and policies about their economy are derived from the analysis of these random sampling data.

There are also population and economic censuses in China. Although each census has different legal backgrounds or motivations, the ultimate goal is similar in order to provide relevant, timely and reliable data for research, analysis and support of final decision-making. As a fast-growing economic power, China's statistics are increasingly valued and have great influence on the world, but they face many of the same challenges.

Using advanced technology to start statistics 2.0 times

Science and Technology daily: In the statistical application, you think the arrival of the big data age to the statistical application of what kind of opportunity?

Hu Shanqing: We know that, since 2000, the ability to capture and store large amounts of electronic data has risen, and new methods have been discovered and broadened, and the big data age has come. A large number of global data are electronically generated and large data outbreaks grow to stimulate and generate more demand for more timely and broader information.

But the collection of large data is almost never designed according to probability, and usually has no structure, and can not carry out traditional statistical analysis. Large data and advanced technology provide an ideal time to start Statistics 2.0.

The visualization and processing of complex data must become the core of Statistic 2.0, and the statistical method is used to tell stories. The dynamic framework retains the original functionality of the traditional framework, captures the latest data in time, facilitates real-time analysis, and enables flexible development of dynamic frameworks to promote innovative practices and innovative products.

If the cost is right and can be presented as a random sample of high efficiency and high quality data analysis, is there any reason why we should not study the whole population? It should be noted, however, that the so-called "big" in the data is a relative concept, based on the percentage of sample size in general, not the absolute concept of data storage.

Sci-tech Daily: So how do you understand the previous statistical stages? That's statistics 1.0.

Hu Shanqing: The concept of random sampling was introduced for the first time in the late 19th century and was fully accepted as a science subject after more than 40 years of controversy. Some people slowly began to pay attention to random phenomena, and gradually focus on the analysis of random probability collected by the representative data, so that statistics become professional, mathematical statistics so rapid development.

Mathematical statistics with probability as the professional theoretical basis, to reach the international consensus of the Standard, open Statistics 1.0, and become an epoch-making watershed. But the traditional census and random survey is static and timed, it is impossible to meet the dynamic demand of expanding without any basic change.

The 21st century statistical system and methods are characterized by the exquisite application of a large number of longitudinal data, the combination of multiple data sources, rapid and simple delivery of information, at the same time can continue to strictly protect the stability of and data security, and certification of accuracy and reliability.

It is more appropriate to compare large data to "sand"

Science and Technology Daily: The industry's discussion of large data is getting hotter, and there is even a tendency to have large numbers of myths, do you think the big data is sand or sands?

Hu Shanqing: Many people will compare large data to the sands, even think can be readily available, if so, many people have already made a big fortune. It would be more appropriate to compare large data with sediment, but a large amount of sediment itself would be valuable. The ancient idea of a heap of sand into a tower, today is the raw element of silicon, the higher value is that it is not fully developed a new knowledge. This large pile of silt, sometimes contains some sands, sometimes there is no, very few occasions, there will be many. In any case, investment effort, digging for sand, no software or hardware can be replaced.

Science and Technology Daily: data for the construction of smart cities is also important, in the big data age, how do you think of the City of Wisdom in China's landing situation?

Hu Shanqing: China promotes smarter cities, bringing the original national base database to more manageable municipal systems and then to the provincial or larger areas. By the end of 2013, China had set up at least 193 intelligent cities to pilot the city. There is no doubt that in the next few years China will continue to promote the construction of smarter cities.

I understand that, October 29, 2013, China's first urban public Information service platform released, the platform has been launched in several intelligent city pilot implementation. It provides a one-stop centre that serves millions of citizens and breaks the household register, enabling the public to use a secure smart card as an additional channel to achieve the previously separate urban function.

Pioneering this system is a modest beginning, it represents the actual work in progress, helps lay the groundwork, and constructs successful Chinese urban informatics and applications.

At a seminar in Chicago at the University of Illinois, we also made a paper proposal for a Chinese city through big data and statistics, and now the proposal has been contested from 90 proposals to participate in the second meeting in August.

(Responsible editor: Mengyishan)