I have finally had a chance to sit down with this webinar, than you so much for making it available! I have a question on bit where you were talking about silica gradients (around 1:17:36) and how one can use a steep gradient (e.g. an abrupt change in Si29 values) to select the data that isn't changing rapidly. I wasn't clear from that quick mention exactly why you wanted to do that in the example case, just that you can.
However, the comment made me wonder if this approach would be useful for imaging--specifically to locate and reject "mixed data". It often happens when an opaque mineral (sulfide or oxide) is in contact with a silicate matrix mineral that the data points from the border between the two often wind up plotting somewhere between the composition each of the two phases. If this would be useful for filtering out mixed data in imaging experiments, what approach would one use to be able to benefit from it?
Thanks,
--Riia
(With apologies for the vagueness of the question--this is me brainstorming directly after hearing you mention the topic, well before I have tried to apply any of the lessons in the webinar)