Statistical techniques of multivariate analysis applied to the interpretation of climate change variables

Abstract

Multivariate data analysis are a very useful tool in data series with a large number of variables, which often do not have a direct correlation, but which need to be interpreted and estimated. An example is all the data that may be related to climate change. Countries make measurements of many factors that can be cause or are a consequence of it. This provides very large databases, which are difficult to interpret. Analysis methods as Principal Component or Factor Analysis help the interpretation and grouping large number of parameters in simpler series. For this study, data from the World Bank were used, specifically for Latin American countries. Data were selected on agricultural land, forest area, protected land areas, population growth, total population, urban population growth and urban population. All of these seem to have some correlation, but the same is not so obvious and especially when it comes to measurements in different units. However, with Principal component method, we found groups that could be related to facts like the need for food, the need for land for housing and the loss of ecosystems. In the case of Factor Analysis, the groups in the factors found show concepts such as land use, total populations and population growth. In both analyzes the usefulness of these methods for the interpretation of large groups of data is evidenced.