So removing data products removes peaks! That's good. So these peaks really correspond to the data products, and the large one corresponds to the "clear" data product.
What can we do about the "clear" data product. How might we get it down to compliance level?