Multi-source macro data process based on the idea of sample=overall in big data

The idea of sample=overall in big data is very suitable for multi-source macroeconomic data processing compared to a selected sample, and is proved to strengthen the role of dynamic networks and dynamic system stability.

Abstract

For the current applicable discussions on the idea of sample=overall in big data processing, this paper selects macro data from multi-source including influence factors of smart city from 17 districts and counties of Shanghai as an overall sample, and standardizes the data. Then, another sample is created from output of Principal Components Analysis (PCA). By comparing the two types of samples in the study of cluster analysis, three conclusions are founded. Firstly, standardization of data processing can serve to strengthen the role of dynamic networks and dynamic system stability. Secondly, factors beyond the principal components also have information carrying capacity and the impact capacity to complex dynamic systems. Thirdly, the amount of information carried by the non-principal components in practical application is much larger than the amount in measurement. Thus, we prove the idea of sample=overall in big data is very suitable for multi-source macroeconomic data processing compared to a selected sample.