And moreover, the author’s conclusion is that the dataset is bad.

> While I had expected some cleanliness issues, I wasn’t expecting to see this level of quality problems from a dataset that I’ve seen referenced pretty extensively across the web

Yeah, but they're still providing a dataset that's just plain bad. It's hardly relevant how many sites link to some other site, if it's dead.

It's only bad data if it does not include what it claims to include.

If the dataset is defined as inlinks, and it is inlinks, then the data is good.

part of the problem is it's not the number of links, it's referring subnets. Fairly certain this includes, script tags.