You must log in or # to comment.
Something’s odd with the numbers from fediverse observer. Numbers shown in monthly graphs should be about 30 times higher than numbers shown in daily graphs, but they are about the same
I always assumed its daily reporting of monthly active users. the nodeinfo format that is used for reporting the data does not show daily active users iirc (it is seemingly capable of reporting weekly users).
Its probably the average number per day in a month(the total number divided by 30)
Thanksgiving?
Sorry. That was me. I ate them all.
I hate the fact that the vertical axis starts at 7 million, making the drop seem deceptively large
Seems fine to me, the axes are easy to understand and there would be a lot of unnecessary whitespace otherwise. Though, it does require some reading comprehension, and that one actually looks at it and not just skims over.
This reading comprehension joke it overused and it doesn’t even make sense here. It’s well-known that you need at least one of those little zigzag indicators when the graph doesn’t start at 0 in most cases to avoid people misinterpreting the graph and to make it much more visually clear.
I was of the impression that reading a graph also required understanding of regular writing/reading, but I’m no native speaker, so I’ll gladly stand corrected.
I’m not sure what you mean by “one of those little zigzag indicators”, do you perhaps mean leap/break in data denoted by the “Squiggle”? I don’t think any data below 7m is included in this graph, so, if I understand you correctly, then that wouldn’t be a proper use of said squiggle.
Yes, I mean an axis break to denote that part of the axis has been omitted. Using it to show a gap between 0 and the lowest included data in the dataset is very common and is a proper use even according to your own source:
A zigzag on the line of the x- or y-axis in a line or a bar graph indicating that the data being displayed does not include all of the values that exist on the number line being used
Beginning the Y axis at 0 and using an axis break to go from there to 7 million allows you to see the same amount of detail as you can in the OP, while visually signalling to the reader that the scale on that axis does not show a full 0 to 10 million range. This increases the chance that they’ll read the graph correctly. You can justifiably blame someone for reading it wrong, but the point of a graph is to communicate, so minimizing the chance for misinterpretation is a good idea.
I agree that this way of displaying the data is appropriate, but it would be nice to have a very visible indicator of this. Some kind of highlighted “fold” line or something at the very bottom of the chart, maybe. If I can deduce the units from context, and the trend is more interesting than absolute numbers, then I’m not going to look at the axes most of the time
Using compressed axes to display data was literally “How to identify misleading statistics 101” in middle school for us…
It seems fine to you but for the majority of people it’s misleading most people look at the lines and the relative distance between them to make judgment calls. Not literally the entire point of graphs, to visually display information.
This is a well-known effect and is taught in pretty much every major curriculum.
And the above was literally how I was thought to represent data in university. Maximize the areas of interest, make sure to properly label your axes (lest they become misleading), and remember to trim empty space where relevant.
But it appears that proper graphs for science and engineering reports may not be used for representing data to the common man, as it must be assumed that, even for the most simple of graphs, the common man will only look at the funny line, but not the graph itself.
Seems like a major server not reporting their numbers.
A 20% drop in use on American Thanksgiving doesn’t seem even remotely outlandish to me
I still doubt we make 9 millions posts per day, there is another graph dedicated to comments, which has a much tinier drop on Thanksgiving: https://lemmy.fediverse.observer/dailystats
I’m pretty sure ThePicardManuever posts 9 million times a day.
this looks more like total posts tbh
Well, yeah. Posters actually put in effort. Commenters are lazy people who just shit out their terrible opinions.
Yea well, that’s just like, your opinion, man
I dunno if people would post so much for upvotes alone. Comments are nice, and I get friendly ones pretty frequently on my posts, and try to leave just as many on others posts :)
Edit: wait, I think I just wooshed myself
It was mostly curmudgeonly humor, but really it does take less effort in general to make comments than it does to make posts. So it would make sense that it would see less of a dip when people are busy. It’s easy to just jump on and make a few comments or participate in a discussion than to put the effort into making a post of some kind.
It really depends on the kind of person, though. I’d imagine some people find it easier to provide content than to comment on it or respond to others doing so.
Hey, my opinions are awesome but you are correct about the shitting and laziness.
Commenters are the worst.
Especially the ones that have custom profiles with names and shit.
Like Bro you aren’t fooling anyone, we all know you aren’t really a squid that flies.
For sure; I sure wasn’t on my phone at all that day, and I’m usually on Lemmy daily. Too busy cooking and mingling with family.
Likewise. The following day as well for me. Oh, and the day after that too. I didn’t return to my regular routine until today.
It’s Thanksgiving. Either a massive drop in posters due to travel or family or whoever was sending the data was similarly out.
I had definitely noticed things were more quiet than usual on English Lemmy, so I figured it was just the Americans being busy with that.
I still doubt we make 9 millions posts per day, there is another graph dedicated to comments, which has a much tinier drop on Thanksgiving: https://lemmy.fediverse.observer/dailystats
That makes a lot of sense.
Possibly an outage from a patch no?
It’s Thanksgiving. A lot of people visiting with friends/family. Idk how to access the numbers but I’d imagine last year was about the same
I still doubt we make 9 millions posts per day, there is another graph dedicated to comments, which has a much tinier drop on Thanksgiving: https://lemmy.fediverse.observer/dailystats
This should say total posts, not average, right?
I am thinking it is probably comments and not posts as well.
There’s a separate graph on that page for comments, which are up around 17 million.
10 million comments by 40k or so active daily users seems incredibly high.
This is showing the total number of posts, not just new ones. The title is really misleading…
Is it? That would make more sense. And then, the number of posts would go down if a very large instance would not report for a day (huge instance in this case, we are talking an instance with 2 millions posts out of 9)?
Probably just a reporting bug. Comments stayed consistent.
Oh, definitely
We need to post harder, not smarter
What would cause this?
Assuming it’s not a bug, maybe Americans busy with Thanksgiving? I know I was on less as a result
I’m thinking that, a lot of users were hanging out with relatives and traveling.
Yeah, I barely looked at Lemmy Thursday - Saturday.
@The_Picard_Maneuver@lemmy.world must have taken a Thanksgiving break lmao 😂
Haha, I did. I still posted a few, but I was out of town with family for days and was barely on my phone.
som kind of bot going down maybe?
Turkey day travel and spending time with family by prolific posters is likely.
People preparing xmas gifts and planning for holidays?
That looks much more like a data artifact than an accurate representation of behavior.
I think that the trajectory of the three low dots matches the overall slope very closely in a way that looks far more like a flat subtraction of all three. If it was behavioral, I think you’d see the behavior come and go over the course of several days.
Maybe it was some sort of federation issue?
You are right, that is quite a dramatic drop and rise.
