Some patent records affected by a 100K cut-off
We've been looking into why some really well-known chemicals seem to have no patents since 2014 ... and it turns out that ~1400 chemicals are not imported from google because they have >100K entries and this would expand the dataset way beyond a manageable size. Current examples include the individual xylene isomers (o-, m-, p-).
We are looking into possibilities .. perhaps for now if you don't have (m)any patents since 2014 you could print a warning that the dataset may be incomplete? (or put a "last patent count" date in the plot)? Most examples seem to stop at 2014, but some have one or two values in 2015/16 - this is data that comes from different sources (IBM, SCRIPDB)