Good analysis! This is the kind of data that I find fascinating. A dashboard is a great idea but even updates about these figures would be helpful. Hivewatchers keeps an eye out for SPAM, but I have spent a lot of time wondering how we can examine and verify some of the content that has already been created. The first 6 days are key for upvotes, but most upvotes are usually made within the first hour of being created.
For some more stat suggestions;
- how many comments are posts that are !PIZZA calls? (good stats, user engagement)
- how many stories/comments have NOT been upvoted on? (separate spam/possible treasure)
- how many users have posted/commented within the last hour?
I honestly have a lot more ideas.
I see other good suggestions in the other comments as well.
Thanks for sharing that link! I watched it for maybe a little bit too long, it was a pretty nice implementation of sitting back and watching the world go by.
I have some logic written to detect any calls to bots that have an exclamation mark in front of them. I think some sort of hybrid metric that looks at comment length and complexity (of the language used) will find those that aren't just calling the bot for people to farm the hive engine tokens.
I think the other interesting thing (if it is there in the data) - I can probably work on this as a goal - is ... what was the oldest post that got a comment that week?
Will require me to do one big pull of data from hivesql, or maybe download and interrogate the block log for historical data, but it would definitely be very interesting!
That IS an interesting link for sure! Thanks for sharing!
What is the Podping thing btw?
I believe https://3speak.tv/ uses Podping. @brianoflondon knows about this.
"Podping leverages Hive to solve the RSS polling problem."
Recent post --> https://peakd.com/podping/@brianoflondon/podping-on-hive-is-quietly-announcing-3000-podcast-episodes-per-hour
Cool! Had never heard about it, but saw it popping up. Thanks for explaining! ;-)
!INDEED