essentially you have a list of buckets, accessible with aggregations['2'].buckets (btw, I would recommend renaming "1", "2", and "3" into something more meaningful). Each bucket has a field datum.key a sub-list datum['3'].buckets which needs to be flattened. Inside each sub-bucket you have two fields -- datum.key and datum['1'].value. Once you flatten everything into a 3 columns, it should be fairly trivial to plot anything you need. Hope this helps!