4 ms·
Note that with the Hacker News dataset now on BigQuery, downloading the dataset is unnecessary and the analysis in the post can be done in seconds: https://news
by minimaxir 11y ago
Note that with the Hacker News dataset now on BigQuery, downloading the dataset is unnecessary and the analysis in the post can be done in seconds: https://news.ycombinator.com/item?id=10440502 https://news.ycombinator.com/item?id=10440502 (although, parsing the domains using that particular dataset is harder since it requires a regex to extract the domain from a URL)
As a side note, I'm sorta annoyed that a data post with so few data and visualizations, but with a linkbait title and selfpromotional startup copy, was far more successful than any of my data posts. Did I miss the part where the OP explained how he used math / how he ranked the Top 50 startups?
- maxpupmax 11y agoI think what explains the difference is that the author is a writer by trade -- not a statistician. The author successful tells a simple story with unexpected tidbits and concrete language (eg "every submission you see here is certifiably timeless") and does a good job following writing best practices like "not burying the lead". You might be interested in reading "Make it Stick" (it's where I learned about this subject from) if you haven't already -- and/or the Priceonomics guide to content marketing. Clickbait title probably didn't hurt either.
- minimaxir 11y agoThe point is that is it not a good thing, especially in the case I'm having difficulty reproducing the results.
- maxpupmax 11y agoI'm not qualified to comment on what's "good" or not. I'd just be willing to bet that if you wrote your own best-of with better/reproducible methodology AND borrowing some established copywriting practices you'd easily outrank this post... but only if you do both parts.