Open
Description
It seems that the examples gallery auto-generates thumbnails by grabbing the last figure in a notebook. However, this has often resulted in thumbnails that are either un-informative or even mis-informative. For example, the thumbnail for GP-circular is actually a contrasting example the author provided of how not to make a circular GP. In a similar vein, the "Mauna Loa continued" thumbnail is not particularly informative.
Perhaps the notebooks could be tweaked to simply put a more useful figure last, or thumbnails could be specified via a cell metadata tag in the notebooks, or a thumbnail could be uploaded separately from the notebook?