The opportunity of data mining for macroeconomic data analysis: A case analysis of East Java Province

The opportunity of data mining for macroeconomic data analysis: A case analysis of East Java Province

macroeconomics, data mining, CRISP-DM, cluster, East Java

The conventional data analysis in economics is based on a model derived from economic theories. In contrast, data mining is a data-driven analysis to extract data and find a pattern describing the empirical interaction between variables. The emerging area of data mining offers an opportunity for extracting information from macroeconomic data. However, it is still a challenge for economic researchers and policymakers to embrace data mining because it is closely related to the information technology discipline. This study responds to the limited use of data mining in the economic area by analyzing macroeconomic indicators published by the Indonesian Central Bureau of Statistics. The primary purpose of this study is to offer a case for using the data mining approach for macroeconomic indicators. The specific objectives were (1) to introduce the Cross-Industry Standard Process for Data Mining (CRISP-DM) as a process framework and Knime Analytics Platform as a data mining software for macroeconomic data analysis; and (2) to characterize East Java regencies/municipalities based on their macroeconomic indicators and region profiles. This study was categorized as secondary and quantitative research. The unit of analysis was the regency/municipality. Five macroeconomic indicators: Human Development Index (HDI), Gross Regional Domestic Products (GRDP), poverty rate, Gini Ratio, and open unemployment rate, were selected as the variables. Four region profiles: area, population, population density, and the number of villages were included in the analysis. The clustering model was implemented through Knime’s workflow. The result of clustering grouped 38 regions into three. Its applicability and simplicity indicated the appropriateness of the CRISP-DM process framework for analyzing the structured official data. Furthermore, the predictive model, applied to past years’ datasets, revealed the regions that experienced improvement and shifted their membership between clusters over three years. Moreover, the inclusion of region profiles has provided a better understanding of underlying factors explaining the association between macroeconomic indicators. This study suggests that the East Java Government considers different facilitation-focused programs based on the characteristics of three clusters for better budget efficiency. This research adds to the literature on economic development, particularly by introducing data mining, the CRISP-DM method, and Knime software to analyze macroeconomic indicators of regency/municipality

Gunawan

perpustakaan horizon karawang

2021

fajar bagus w

pdf

english

text