This is the third of three articles on an analysis of Hacker News community, utilizing HNProfile.com – a demonstration of the Metacortex platform. The first article is Trends on Hacker News: Politics and the second is Trends on Hacker News: Programming Languages & Tools.
This article primarily focuses on the activity of the Hacker News community, when it’s most active and the kind of people who are active on Hacker News.
When is Hacker News Most Active?
Activity on Hacker News is extremely consistent — primarily due to it’s large, international user base:
If you can read the pattern, somewhat difficult on this chart, it always peaks Wednesday and roughly half as much activity on Saturdays (Sunday being a close second). This also matches exactly with the trends identified in email data. There are also holidays which can cause some deviation, but in general — it’s consistent. Typically, this means for the most engagement you should post an article Wednesdays.
In terms of activity on an average day, it fluctuates somewhat between weekdays and weekends. However, as I’m writing this on Saturday you can see both the average overall and what it looks like today:
The time is represented in central standard time (CST), and pretty much matches expectations. The early morning until roughly noon (CST) is relatively slow and then picks up in the afternoon. In terms of time, this also pretty much matches the work day in the bay area. 11 am CST is 9 am PST, so from 9 am PST until 9 pm PST is the most active. Makes sense to me.
When is Hacker News in the Best Mood?
In terms of mood, the Hacker News community is also extremely consistent. The sentiment is always on average around what we classify at Metacortex as slightly positive:
Upon review, even the average hourly sentiment, is extremely consistent, right around a sentiment score of slightly positive; a sentiment score of 0.2 out of a score that ranges -1 (negative) to +1 (positive):
Generally, this makes the Hacker News community one of the most positive communities I’ve found. Reddit, for instance, is significantly less positive.
Hacker News Community Size & Growth
Warning: This section is slightly less accurate than other sections. Unfortunately, I am missing data from mid-2015 through 2016. In 2017, I began collecting data utilizing a different method, unfortunately, the system did not have 100% uptime and may impact results.
In terms of community size and growth, Hacker News has close to 300,000 “active” accounts, likely this means much less actual users (I discuss in this in the conclusion). The growth looked extremely consistent, then appeared to pretty much level off. As mentioned above, this could just be due to the system change above. On the other hand, it could be just reaching the market saturation.
For 2018, I’d expect roughly two million comments on Hacker News (by year-end):
As you can see in the graph above, the missing data is somewhat of a factor. At some point, I suspect I’ll fill in the missing data, however, today this is what I have (and I don’t have the time to backfill the data). Regardless, it does appear the comments have plateaued, perhaps reaching all of the “hacker” community.
Alongside the growth of any community comes rules and regulations and Hacker News is no different. However, their moderation seems relatively light by some standards. Utilizing HNProfile.com and searching for YCombinator, it’s possible to identify the moderators of Hacker News:
What’s interesting, is it’s also possible to see a gap in time where neither of the moderators are moderating, that would be right between 9 pm to midnight:
So, if you’d like to post content that won’t be moderated, by YCombinator that’s the time! However, that’s not to say the community won’t moderate you… It turns out, with further investigation, it appears there are many community members who take it upon themselves to remove duplicates. It does however, appear they have some admin privileges and are empowered by the official moderators to remove duplicates (or at least notify them quickly).
Expertise on Hacker News
One thing I think is also important to note, is there are experts in virtually any / every field. How to quantify that is a bit hard, there are quite as many wonderful graphs. However, when you search for experts in Docker, for instance, the creator of docker is the first results:
From basic testing, this works for Docker, AWS, Keybase, TensorFlow, Postgres, GCP, Lisp, Golang, Chrome, and the list goes on… Almost all of them have the first result as either the creator, maintainer, or lead developer on the given topic.
Personally, this leads me to believe the Hacker News community may be one of the largest “expert networks” in the world.
Finally, one thing I’d like to mention – there appear to be many duplicate Hacker News accounts, per real user. According to Metacortex, there are only ~90,000 unique “author profiles” (i.e. real users) which posted on Hacker News since the beginning of 2018. An “author profile” is a single real identity, which could represent multiple accounts, even cross-domain (for instance on reddit.com)
What that means: of those 300,000 “active” accounts, ~200,000 accounts are either:
- Throw away accounts
- One of multiple accounts owned by a single real user
I suspect it’s relatively common for users to either post from an alternate account or use a throwaway. For instance, I know I personally have two Hacker News accounts.
Now, you may be asking yourself: “Wait, how do you identify an Author Profile?” How we identify “Author Profiles” is our secret, however needless to say it does generally work (provided enough comments are made on each account):
Overall, it’s been interesting and enlightening to review the Hacker News community. If you have not read my previous two articles, I recommend it:
Personally, I find the most interesting aspect developing HNProfile.com has been able to identify experts and identify which topics are growing over time. This has worked out handsomely when I needed assistance or need to identify what to invest time into.
If you are interested in utilizing Metacortex within any organization, please reach out! You just send content into the system, and from there a knowledge graph is constructed, on which all this analysis is automatically done. It typically takes less than a day to integrate into any organization.