Page 373 - Special Topic Session (STS) - Volume 2
P. 373
STS500 Li J.
this paper aims to explore how to deepen the application of big data in
government statistics.
In the qualitative analysis method, it is required to state the thinking of
big data and the differences between it and the government statistics on the
basis of recognizing the concept of big data; to summarize the application
characteristics of big data in enterprises so as to the thinking for the
application method of big data for government statistics; to qualitatively the
available room for improvement in each link of government statistics
process, including data collection, data analysis and data issuing, so as to
pertinently put forward how to use the big data to improve the current
situation.
3. Result and conclusion
(I) Stating the differences between the thinking contained in big data
and the government statistics
Besides the concept and characteristics of big data, the differences
between the thinking contained in big data and the government statistics
should be also disclosed. The big data represents the data-driven thinking
instead of any preset hypothesis. The data should speak for itself. In other
words, after the data size reaches a certain extent, the regular and trend-
oriented information will appear, and such information is always unexpected.
The government statistics represents a thinking of hypothesis verification. In
other words, the sample data is collected and analyzed to infer the overall
situations or verify the original hypothesis. The differences between them are
shown as follows: (1) the thinking of big data has higher tolerance for
imprecision, in which the precision requirement for individual data is relatively
low but the trend manifested after all the data is gathered together is
emphasized; (2) the thinking of big data expands the application range of
simple algorithm, and the big data has a lower dependence on algorithm if
compared with the government statistics; (3) different processing method for
“useless” data, the government statistics is used to eliminate or reduce the
general impact of “useless data” on inference by replacing or expanding the
sample data and optimizing the algorithmic routine, but it is thought in the
thinking of big data that each kind of data is provided with valuable
information and the “useless” data is utilized by means of reverse thinking.
(II) Summarizing the application characteristics of big data in
enterprises
The application characteristics of big data in enterprises are shown as
follows:
(1) mining and analysis of massive data (i.e., “sample=population),
indicating that new information may appear after the data size reaches a
362 | I S I W S C 2 0 1 9