21–28 of 28 posts
"High quality comments" for a user in general is completely subjective. They might have "high quality comments" on C programming, and "low quality comments" on marriage equality. You'd need a custom filter for every possible topic of conversation, and even then you'd have to filter based on what you find to be "high quality". It's easier to just filter based on popular opinions or popular people, which is what karma…
"Squawk to Talk" ratio Squawk = number of words in your hacker news comments Talk = number of words in children of your hacker news comments
I don't think comment karma is publicly available through the site or API, but average comment karma could be calculated based on total karma of user minus sum of post karma divided by number of comments.
Example: https://hn.algolia.com/api/v1/search?query=bar&tags=comment
Whatever metric you use, since HN as a community is effectively training the metric, you'll simply capture what HN users consider as "high-quality" comments. HN users do have their biases (in my opinion): - Anti big company (GOOG, APPL, ... Lees so for APPL thouhg) - Mostly left leaning - Value political correctness over accuracy - Anti defense and government (loathe 3 letter agencies, despite being fully cognizant o…
Are you by chance right-wing? I'm solidly left wing and personally more often see the opposite bias permeate HN comments for the various topics you listed.
I don't think it's possible. A high-voted comment isn't necessarily high-quality, just really popular. I recently had a command get to 36 points [0]. I'm honestly not convinced of it's quality, even though it's my own comment. It was as a strongly worded opinion based on a high level of cynicism, not research. But people liked it, so it got a decent number of votes. [0] https://news.ycombinator.com/item?id=19060284