3 ms·
Relevant: if you're looking for a way to scrape all Hacker News submissions and comments, I have a GitHub repository containing Python scripts showing how to do
by minimaxir 11y ago
Relevant: if you're looking for a way to scrape all Hacker News submissions and comments, I have a GitHub repository containing Python scripts showing how to do just that: https://github.com/minimaxir/get-all-hacker-news-submissions-comments https://github.com/minimaxir/get-all-hacker-news-submissions...
With respect to the article, I've done data analysis on Hacker News myself (e.g. http://minimaxir.com/2014/10/hn-comments-about-comments/ http://minimaxir.com/2014/10/hn-comments-about-comments/ ) and I'm a little skeptical. I was not a fan of using NTLK to classify positive/negative because the models used for training sentiment are based on reviews, not internet site comments, and when I had used NTLK, the results were skewed significantly. I used an alternate model mentioned in the linked post which IMO worked better.
The examples used in the article seem more like correlation-implying-causation and cherry-picking. There are many, many counterexamples of startups people loved on HN which died a painful death.