Excellent point about the batch sizes--we have no clear way of knowing what they are and can only guess from what we see across VAERS. This distribution of adverse event counts is highly skewed. I'm also trying to reproduce his analysis but no luck so far. The VAERS data are split across multiple files, and I presume the variable VAX_LOT is what he is using to assign batches. Another key question is: How is batch date assigned? The dates form his x-axis which underlies pretty much everything he concludes. We have VAX_DATE for each row (VAERS_ID), but these vary for each VAX_LOT. I tried both min date and mean date but am not getting plots anything close to what he showed. I am including the non-domestic data to get the fullest picture.