Gain an overview of global co-variability patterns in 2m air temperature (t2m), zonal and meriodional surface wind speeds (u10 and v10).
The ERA Interim data set is a homogenous re-analysis dataset with global geographic coverage. It is thus ideal for your purposes. You will need files for the above-mentioned variables.
Apply a hierarchical cluster analysis to annual means of your climate variables, decide on an optimal number of clusters (a cut-off number in the procedure) and plot your results.
Hint: If you plot your cluster results for several steps in your clustering, you can visualise and track the categorisation process. This will help you understand the method, allow you to make a more informed choice of your cut-off cluster number, and allow you to explain this better to an audience.
Optional: Apply k-means clustering corrections to your results, i.e. use centroids from your hierarchical clustering as input for a k-means clustering procedure.
Optional: Repeat the analysis for a specific extratropical region.