Finding nonlinear relationships between the Sustainable Development Goals and climate change with partial distance correlations
This website allows you to explore the results of the publication Complex interlinkages, key objectives, and nexuses among the Sustainable Development Goals and climate change: a network analysis that is freely available at The Lancet Planetary Health.
Given the large analysis over 35 country groupings, most of the results are found in the Supplementary Material of this publication. This website is intended to make the search for results of specific country groupings more accessible.
The two main differences of our approach to existing studies are the consideration of non-linear (non-monotonic) relationships, how the following animated Figure 1 explains, and the consideration that any dependence (or correlation) between two variables may originate from a third variable, that is not taken into account initially. These third variables are called confounders or lurking variables and make the dependence between the two variables spurious. We try to explain this in Figure 2 below.
Figure 1:
Figure 2:
See our Google Scholar page to find out how to cite this research.
Download the numerical results of all country groupings as csv files.
Explanation to downloaded tables: You will find six columns. The first column is an index and can be ignored; the second (“pair_0”) and third columns (“pair_1”) are the SDG numbers (“1” to “17”) and climate change (“T”); the fourth column (“min_dcor”) is our measure of strength of an interlinkage, the minimum partial distance correlation; the fifth column is the p-value of the statistical test for partial independence (we chose a confidence level of 5%); and the sixth column is the conditional set of variables (SDGs + climate change) that resulted in the minimum partial distance correlation.
For any questions, comments, or improvement suggestions, please do not hesitate to contact me at fjl1218@ic.ac.uk.