Data mining applications with logistics indicators of OECD countries

Today's technological developments cause the accumulation of billions of big data, especially in the fields of economy and logistics. It is very important to analyse the accumulated big data using special software for data mining so that it reveals important information for researchers and policymakers. The aim of this study is to analyse the big data accumulated in the economic and logistics fields of OECD countries and to make comprehensive policy inferences. In line with the purpose of the study, the analyses applied in WEKA 3.8.6 software and STATA 16 package program were carried out in four steps. First of all, cluster analysis was carried out separately using the k-means algorithm for the GDP, export, import, inflation, and unemployment rate variables of 29 OECD member countries for the period 2013-2022. As a result of the analysis, countries were divided into four clusters. Germany and the USA were included in the cluster of the most developed countries, while Belgium, France, Japan, Canada, Korea and the United Kingdom were determined to be developed countries. Greece and Spain represent developing countries, while Turkey has become a developing country, particularly in recent years. Estonia, Latvia, and Lithuania are included in the cluster representing less developed countries. In the second stage of the analysis, logistic indicators that cause differences between countries were determined using the 2019 data of 20 transportation variables of OECD countries and the J48 algorithm of decision trees, one of the classifier methods. It has been determined that the transportation variables that cause the differences in the economic development levels of countries are railway infrastructure investment, total domestic transport infrastructure investment per GDP, railway infrastructure investment per capita and the share of CO2 emissions from international aviation bunkers in total CO2 emissions. Thirdly, the 2025 forecasts of these four transportation variables, which were found to differentiate the economic development levels of OECD countries, were made separately for each country using artificial neural networks. In the last step of the analysis, the effects of railway infrastructure investments, railway infrastructure investments per capita, total domestic transport infrastructure investments per GDP, and the share of CO2 emissions from international aviation bunkers in total CO2 emissions on economic growth were examined using the Driscoll-Kraay estimator. The analysis was carried out by using annual data of OECD countries covering the period 2010-2019. According to result of the analysis, it was determined that the effect of railway infrastructure investment and per capita railway infrastructure investment on GDP was statistically insignificant, while the effect of total domestic transport infrastructure investment per GDP and the share of CO2 emissions from international aviation bunkers in total CO2 emissions on GDP was statistically significant.