Getting this data is not easy. Many companies that make 3D printers are either private companies that do not report results or are larger companies where 3D printers are one of many products they manufacture. Some companies study this information through mining public sources and surveying users for their opinions and experience. The result of some of these studies are available for a fee.
Occasionally, a trade publication will survey data sources and produce an article. In other cases, a trade pub will publish an article generously offered by a commercial contributor.
It is always difficult to know what is true when abstracting information from obscured, noisy, and biased information sources.
Your question itself includes a bias. You use a words that include a value judgement: "but that likely doesn't translate to sales, at least for the more overpriced ones."
The article you reference is not a deeply researched investigative piece. It is simply some product details for the five printers in 2016 which sold the most on Amazon.com. It doesn't include printers which were not sold on Amazon, so it leaves out any printers which use a different distribution channel. Also, the article include an link, probably which generate revenue back to the magazine, to each of the five printers sold through Amazon.
To summarize, it is very difficult to aggregate this kind of information. Those who try to do so like to be compensated. A list of the top five devices on Amazon is a biased list.