Multivariable statistical modelling on panel data

Panel data consists of observations over units over a certain period of time. However, the inclusion of time series in the structure of panel data brings both its features and problems. Panel data analysis shows the change of a unit over time or the change of a variable between units and enables the detection of a relationship by creating a model with the variables. It also provides the opportunity to compare the relationships in terms of each unit for the variables. In this study, observations for carbon dioxide emissions per capita, gross domestic product per capita and renewable energy for the years 1975-2017 are used. After clustering according to the variables, panel data analysis is applied. In the clustering analysis, five clusters are decided by using multivariate panel data clustering methods. The clusters obtained are named as new panel units and panel data analysis is applied. With the fixed effect model, the effects of the units are estimated, and long-term relationships are analyzed. In addition, clusters are formed according to carbon dioxide emissions, renewable energy, GDP and country averages and cointegration relationships are investigated.