ÌÇÐÄÆƽâ°æ

This is how the source conducted their method, translated: 

"In the work with YouTube, we have over the years built a database with a total of 13,200 YouTube channels. We can retrieve the data about the activity from these channels via Youtube's API. The percentage of Swedes among the viewers of a YouTube clip is very difficult to determine for an external party because that data is not publicly available. Therefore, we are forced to manually determine the main language usage of all YouTubers as a starting point for its importance in Sweden. Many YouTubers also have accounts on, for example, Twitter and Facebook. In these cases, we can cross-check data on the percentage of Swedes and, through this, estimate the account's reach in Sweden. If we cannot find information about Swedishness through these methods, we instead make a normal assessment of the account's Swedishness based on an average value. In several cases, we have also contacted those responsible for the YouTube channel and offered them to contribute their own statistics on the number of Swedes they reach. It is then used as a supplementary data point. The corresponding method described for ranking accounts on Twitter and Facebook is also used for YouTube, based on the following six ranking measures: 1) the number of likes 2) the number of likes per post, weighted according to the assumed proportion of Swedes 3) number of comments 4) number of comments, weighted according assumed proportion of Swedes 5) number of views 6) number of views, weighted by assumed proportion of Swedes. We weight a video on shorts the same way as a video in YouTube's regular feed. Just like on Facebook, we don't measure subscribers on YouTube, as that is an increasingly irrelevant measure of a channel's influence."