stats

The Something Awful Forums > Archives > Comedy Goldmine > The Goodmine > stats

Over There: Jun 28, 2013; by Azathoth

Leperflesh posted:

I feel like something could be done to compare site traffic (e.g., https://www.alexa.com/siteinfo/somethingawful.com) to probation rate, and see if we're handing out more, the same, or fewer hours of probation per user visit over time

but lol I'm not paying alexa for premium traffic analysis stats

Maybe Lowtax can with his $11,000 a month.

# ? Oct 3, 2019 20:07

Adbot: ADBOT LOVES YOU

# ? Jun 3, 2024 14:10

Leperflesh: May 17, 2007

well that traffic includes the front page, apparently, and while that surely isn't a huge amount of visits it could skew the historical results and validity of these very very important and valuable statistics

e. there's also this stuff

is there a historical record of registered users over time?

also is the difference between "users logged in" and "registered users logged in" the number of lurkers currently browsing the forums without accounts? Because that's higher than I would have guessed

e. goddamn how have there been zero users banned today, it's like noon pacific wtf are the admins even doing

Leperflesh fucked around with this message at 20:13 on Oct 3, 2019

# ? Oct 3, 2019 20:09

Snazzy Frocks: Mar 31, 2003; Scratchmo

does the Users Total figure include the Total Users Banned figure?

# ? Oct 3, 2019 20:30

pog boyfriend: Jul 2, 2011

sourdough posted:

Not op, but what it should be is:

Each user is a point in n dimensions, where n is the number of different probation lengths (and bans? Dunno) and each user's value is the number they've given out. So you could be like 100 in the 6 hour, 1 in the 24 hour, 10 in the ban, etc. PCA tries to find common patterns in those dimensions. Like, maybe IKs tend to give a ton of 6ers and bans and nothing in between, old admins usually gave week or month or ban and not much shorter, something like that. The pattern that explains the most variance in the data will be principal component 1, the next principal component 2, and now you can take the n original dimensions and visualize them in just those 2 PC dimensions. Let's say "IKs" and "old admins" were the patterns that were most dominant, now each user being plotted would be how similar they were to the IK style of probes (lots of 6ers, lots of bans, not as much else) on the x axis and how similar they were to the old admin style on the y axis. Points close to each other are users that have similar probe patterns, and that fall in a particular spot on the spectrum of "like IKs" and "like old admins."

Uh, those lines shooting off labeled with the probe durations were probably him trying to show how the principal components related to the actual probe durations, but they're pretty unreadable. So it's tough to tell from that figure what the actual principal components mean in terms of patterns of probations.

no i totally got that, it was just really funny how the axes were not labeled so it could be loving anything and we have no idea

# ? Oct 3, 2019 20:55

sourdough: Apr 30, 2012

Important forum stats coming up lol

Here's PCA on the fraction of probations that each mod/IK gave out that were of each duration (only looking at people that gave out more than 20 total probes). I didn't do raw number of each probation, since the data doesn't have how long the person's been a mod.

The biggest distinguishing feature is unsurprisingly 6 hours, with a bunch of mods only ever giving 6ers. The next thing driving variability was mods giving out a lot of 1 day and not many longer vs. mods giving out very few 1 day and many more 3 day/week/month. So it's like "IK vs not IK" for the main difference, then "lenient mod vs harsh mod" for the second difference.

Top 5 lenient mods : IOwnCalculus, Martyoof, pipes!, Joementum, Butt Wizard

Top 5 strict mods: liquorhead, Inspector_666, Fatal, NotShadowStar, MainPaineframe

Here�s what the top two PCs look like. Positive value means they give out proportionality more of that probe than average, negative means proportionality less than average.

Here�s showing what the top 5 and bottom 5 mods that match each PC look like. For example, �high PC1� in filled black is IKs: they give 100% 6 hour probations. �High PC2� in filled blue circles are �lenient� mods, that give a ton of 1 day probes and not many longer ones. �Strict� mods, low PC2, are the reverse: few 1 day probes and many longer ones. I don't know who NotShadowStar is, but they're a dick, almost all of their probations given are evenly split between being a week or 3 days.

Finally, here�s the whole distribution of mods and where they fall on the PC1 �IK or not IK� and PC2 �lenient vs. strict� axes. The distribution is kinda weird because it's a zero sum thing. If you give out a lot of one duration, you can't give out a lot of other durations, because it all has to sum to 100%.

# ? Oct 3, 2019 22:29

Oil of Paris: Feb 13, 2004; 100% DIRTY; Nap Ghost

NERDS

# ? Oct 3, 2019 22:37

Pewdiepie: Oct 31, 2010

Hello OP. I am curious if your stats account for account name changes. Thank you.

# ? Oct 3, 2019 23:20

Over There: Jun 28, 2013; by Azathoth

sourdough posted:

Important forum stats coming up lol

Here's PCA on the fraction of probations that each mod/IK gave out that were of each duration (only looking at people that gave out more than 20 total probes). I didn't do raw number of each probation, since the data doesn't have how long the person's been a mod.

The biggest distinguishing feature is unsurprisingly 6 hours, with a bunch of mods only ever giving 6ers. The next thing driving variability was mods giving out a lot of 1 day and not many longer vs. mods giving out very few 1 day and many more 3 day/week/month. So it's like "IK vs not IK" for the main difference, then "lenient mod vs harsh mod" for the second difference.

Top 5 lenient mods : IOwnCalculus, Martyoof, pipes!, Joementum, Butt Wizard

Top 5 strict mods: liquorhead, Inspector_666, Fatal, NotShadowStar, MainPaineframe

Here�s what the top two PCs look like. Positive value means they give out proportionality more of that probe than average, negative means proportionality less than average.

Here�s showing what the top 5 and bottom 5 mods that match each PC look like. For example, �high PC1� in filled black is IKs: they give 100% 6 hour probations. �High PC2� in filled blue circles are �lenient� mods, that give a ton of 1 day probes and not many longer ones. �Strict� mods, low PC2, are the reverse: few 1 day probes and many longer ones. I don't know who NotShadowStar is, but they're a dick, almost all of their probations given are evenly split between being a week or 3 days.

Finally, here�s the whole distribution of mods and where they fall on the PC1 �IK or not IK� and PC2 �lenient vs. strict� axes. The distribution is kinda weird because it's a zero sum thing. If you give out a lot of one duration, you can't give out a lot of other durations, because it all has to sum to 100%.

I got a C in Statistics but I appreciate the effort.

You should get another av though because the deep fried crab has been making me uncomfortable.

# ? Oct 3, 2019 23:21

Burt Sexual: Jan 26, 2006; by Jeffrey of YOSPOS; Switchblade Switcharoo

sourdough posted:

Important forum stats coming up lol

Here's PCA on the fraction of probations that each mod/IK gave out that were of each duration (only looking at people that gave out more than 20 total probes). I didn't do raw number of each probation, since the data doesn't have how long the person's been a mod.

The biggest distinguishing feature is unsurprisingly 6 hours, with a bunch of mods only ever giving 6ers. The next thing driving variability was mods giving out a lot of 1 day and not many longer vs. mods giving out very few 1 day and many more 3 day/week/month. So it's like "IK vs not IK" for the main difference, then "lenient mod vs harsh mod" for the second difference.

Top 5 lenient mods : IOwnCalculus, Martyoof, pipes!, Joementum, Butt Wizard

Top 5 strict mods: liquorhead, Inspector_666, Fatal, NotShadowStar, MainPaineframe

Here�s what the top two PCs look like. Positive value means they give out proportionality more of that probe than average, negative means proportionality less than average.

Here�s showing what the top 5 and bottom 5 mods that match each PC look like. For example, �high PC1� in filled black is IKs: they give 100% 6 hour probations. �High PC2� in filled blue circles are �lenient� mods, that give a ton of 1 day probes and not many longer ones. �Strict� mods, low PC2, are the reverse: few 1 day probes and many longer ones. I don't know who NotShadowStar is, but they're a dick, almost all of their probations given are evenly split between being a week or 3 days.

Finally, here�s the whole distribution of mods and where they fall on the PC1 �IK or not IK� and PC2 �lenient vs. strict� axes. The distribution is kinda weird because it's a zero sum thing. If you give out a lot of one duration, you can't give out a lot of other durations, because it all has to sum to 100%.

Where�s my circles? I get both complaints I�m too strict and too lenient. I need stats to self correct, man.

# ? Oct 3, 2019 23:33

sourdough: Apr 30, 2012

Burt Sexual posted:

Where�s my circles? I get both complaints I�m too strict and too lenient. I need stats to self correct, man.

You give too many 6ers to be very separated on the harsh/lenient axis, you're sitting at a 0.24 on the x-axis and -0.02 on the y-axis in that last figure. Maybe I should amend "IK vs not IK" to be "IK/modern GBS vs not IK/modern GBS" :v:

So give fewer 6 hours so you look more like a real mod would be my feedback!

I don't think older GBS mods did 6ers too often, louisgod and BrotherJohnathan for example are out at like -0.65 and -0.5 on the x-axis (but both slightly positive on y-axis, actually). I don't remember other older mods specifically to check any more. Oh, aatrek lol, he's at -0.6 x-axis and slightly harsh on the y-axis.

# ? Oct 4, 2019 00:25

BurgerQuest: Mar 17, 2009; by Jeffrey of YOSPOS

sourdough posted:

Important forum stats coming up lol

Here's PCA on the fraction of probations that each mod/IK gave out that were of each duration (only looking at people that gave out more than 20 total probes). I didn't do raw number of each probation, since the data doesn't have how long the person's been a mod.

The biggest distinguishing feature is unsurprisingly 6 hours, with a bunch of mods only ever giving 6ers. The next thing driving variability was mods giving out a lot of 1 day and not many longer vs. mods giving out very few 1 day and many more 3 day/week/month. So it's like "IK vs not IK" for the main difference, then "lenient mod vs harsh mod" for the second difference.

Top 5 lenient mods : IOwnCalculus, Martyoof, pipes!, Joementum, Butt Wizard

Top 5 strict mods: liquorhead, Inspector_666, Fatal, NotShadowStar, MainPaineframe

Here�s what the top two PCs look like. Positive value means they give out proportionality more of that probe than average, negative means proportionality less than average.

Here�s showing what the top 5 and bottom 5 mods that match each PC look like. For example, �high PC1� in filled black is IKs: they give 100% 6 hour probations. �High PC2� in filled blue circles are �lenient� mods, that give a ton of 1 day probes and not many longer ones. �Strict� mods, low PC2, are the reverse: few 1 day probes and many longer ones. I don't know who NotShadowStar is, but they're a dick, almost all of their probations given are evenly split between being a week or 3 days.

Finally, here�s the whole distribution of mods and where they fall on the PC1 �IK or not IK� and PC2 �lenient vs. strict� axes. The distribution is kinda weird because it's a zero sum thing. If you give out a lot of one duration, you can't give out a lot of other durations, because it all has to sum to 100%.

Very Neat!

Schweinhund posted:

I mean could you look up which ban reasons, word for word, are most common. And list like the top 100?

Sure here you go - it's not that interesting sadly.

Trabisnikof posted:

Any desire to add reg date to the dataset? You could do a lot of fun charts on probation type by reg year over time, longest time til first probation, etc

Plenty of desire, no practical way to do it unfortunately.

# ? Oct 4, 2019 00:34

Tetramin: Apr 1, 2006; I'ma buck you up.

How many times was that home alone gif that was popular a couple of years ago used. Lol no I do not have the link but it�s gotta be in the hundreds

# ? Oct 4, 2019 01:06

stab: Feb 12, 2003; To you from failing hands we throw the torch, be yours to hold it high

Son of Thunderbeast posted:

Tableau users: I threw the .csv into a workbook and uploaded the .twbx here:

https://www.dropbox.com/s/1n8xbxox010uxs2/Leper%27s%20Colony.twbx?dl=0

So far I haven't done anythign with the data at all. I haven't touched Tableau Prep yet so it might be a good excuse to get familiar with that

Awww yeah thats the good poo poo plug the advanced stats right into my veins

# ? Oct 4, 2019 01:14

Schweinhund: Oct 23, 2004

BurgerQuest posted:

Sure here you go - it's not that interesting sadly.

Thanks!

# ? Oct 4, 2019 01:29

former glory: Jul 11, 2011

How many mod actions for busting The Krust?

# ? Oct 4, 2019 01:35

Burt Sexual: Jan 26, 2006; by Jeffrey of YOSPOS; Switchblade Switcharoo

A chart that show probation (ncnt and hours) given by subforum. Preferably sorted,by subforum user ncnt.

# ? Oct 4, 2019 02:30

Farmer Crack-Ass: Jan 2, 2001; this is me posting irl

I'm curious what the breakdown is on how the hours of Pittsburgh Fentanyl Cloud's total probation record break out across mods/subforums. Who hates him more, C-SPAM or SAS?

# ? Oct 4, 2019 06:42

Twenty Four: Dec 21, 2008

BurgerQuest posted:

not posts but lowtax ban reasons word count:

Anyone know the context as to why "Tech / TECH" is one of the top words lowtax used for bans?

Also lol that lowtax has his own name ranking fairly high in that ban reason cloud.

# ? Oct 4, 2019 11:54

fauna: Dec 6, 2018; Caught between two worlds...

# ? Oct 4, 2019 12:53

El_Elegante: Jul 3, 2004; by Jeffrey of YOSPOS; Biscuit Hider

Some of the Sheep posted:

I hope you're joking about this, otherwise you're starting to sound pretty pathetic.

I�ve got some bad news for you

# ? Oct 4, 2019 14:10

Schweinhund: Oct 23, 2004

Twenty Four posted:

Anyone know the context as to why "Tech / TECH" is one of the top words lowtax used for bans?

Also lol that lowtax has his own name ranking fairly high in that ban reason cloud.

people putting in tech support requests to get access to offsite torrent sites (that lowtax had nothing to do with). Also some other tech support stuff and things for the tech support subforum

# ? Oct 4, 2019 15:30

Twenty Four: Dec 21, 2008

Schweinhund posted:

people putting in tech support requests to get access to offsite torrent sites (that lowtax had nothing to do with). Also some other tech support stuff and things for the tech support subforum

I guess that makes sense, I figured if it was such a huge number it would have been something more. Maybe there is? Anyhow thanks!

# ? Oct 4, 2019 17:35

Chokes McGee: Aug 7, 2008; This is Urotsuki.

sourdough posted:

Important forum stats coming up lol

Here's PCA on the fraction of probations that each mod/IK gave out that were of each duration (only looking at people that gave out more than 20 total probes). I didn't do raw number of each probation, since the data doesn't have how long the person's been a mod.

The biggest distinguishing feature is unsurprisingly 6 hours, with a bunch of mods only ever giving 6ers. The next thing driving variability was mods giving out a lot of 1 day and not many longer vs. mods giving out very few 1 day and many more 3 day/week/month. So it's like "IK vs not IK" for the main difference, then "lenient mod vs harsh mod" for the second difference.

Top 5 lenient mods : IOwnCalculus, Martyoof, pipes!, Joementum, Butt Wizard

Top 5 strict mods: liquorhead, Inspector_666, Fatal, NotShadowStar, MainPaineframe

Here�s what the top two PCs look like. Positive value means they give out proportionality more of that probe than average, negative means proportionality less than average.

Here�s showing what the top 5 and bottom 5 mods that match each PC look like. For example, �high PC1� in filled black is IKs: they give 100% 6 hour probations. �High PC2� in filled blue circles are �lenient� mods, that give a ton of 1 day probes and not many longer ones. �Strict� mods, low PC2, are the reverse: few 1 day probes and many longer ones. I don't know who NotShadowStar is, but they're a dick, almost all of their probations given are evenly split between being a week or 3 days.

Finally, here�s the whole distribution of mods and where they fall on the PC1 �IK or not IK� and PC2 �lenient vs. strict� axes. The distribution is kinda weird because it's a zero sum thing. If you give out a lot of one duration, you can't give out a lot of other durations, because it all has to sum to 100%.

Joe can actually be thrown out as an outlier. I think he was a mod for less than a month before D&D broke him.

And I mean broke, not just got into a slapfight he didn't like. IIRC someone posted a poo poo ton of MLP porn in one thread and having to edit each of those out individually is what made him go "No thanks."

# ? Oct 4, 2019 18:16

Schweinhund: Oct 23, 2004

Twenty Four posted:

I guess that makes sense, I figured if it was such a huge number it would have been something more. Maybe there is? Anyhow thanks!

well there's like dozens of bans that are "Tech Support Request: Yo i can't get my account to work on [torrent site] can you fix it"

edit : lol (redacted by me, not lowtax) (remember lowtax had nothing to do with these sites)

quote:

Tech Support Request: Tech Support,

I recently signed up for your forums, and did so before the advertised shut down date of the BTB on the first of the new year. When i went to confirm registration this morning the 31st at http://wdma.biz/x/?reg it said that registration was suspended indef and not to contact customer support. This feature is the only reason that I signed up for the forums, and why I did so before the cut off date, hence why I am contacting tech support. If I am not able to access this location (ex: wdma.biz/x/?=1925) as advertised, then I would like a full refund of my purchase price because not all functions that were advertised can be rendered

Thank you for your prompt reply,

[first name last name]

[his college]

3rd Year G.Surgery

[his address]

Cell: [his phone number]

Stat page: [I don't know what this is]

Email: [school email address]

edit2: don't know why "support" doesn't show up in the picture, but I don't see many "tech"'s without "support" after it.

Schweinhund fucked around with this message at 19:03 on Oct 4, 2019

# ? Oct 4, 2019 18:38

Chokes McGee: Aug 7, 2008; This is Urotsuki.

Schweinhund posted:

well there's like dozens of bans that are "Tech Support Request: Yo i can't get my account to work on [torrent site] can you fix it"

I will never get tired of Dear Richard letters

# ? Oct 4, 2019 21:24

Scaramouche: Mar 26, 2001; SPACE FACE! SPACE FACE!

Breakdown of img-timeline disciplinary actions presented in an img-timeline format

# ? Oct 4, 2019 23:22

ass cobra: May 28, 2004; by Azathoth

Been seeing a little bit of a comeback for our old friend the 12 hour probation over the last couple of days. Thanks to this excellent thread no doubt.

# ? Oct 5, 2019 10:24

Burt Sexual: Jan 26, 2006; by Jeffrey of YOSPOS; Switchblade Switcharoo

Was a pemabanned chart made?

# ? Oct 7, 2019 02:47

BurgerQuest: Mar 17, 2009; by Jeffrey of YOSPOS

Don't think so but could do, apologies, have been doing life the last few days. I was hoping all these PRO STATS GOONS would step up but nooooo.

# ? Oct 7, 2019 08:40

Wark Say: Feb 22, 2013; by Fluffdaddy

Take your time, it's been fun to see diverse stats for this place!

# ? Oct 7, 2019 16:57

Son of Thunderbeast: Sep 21, 2002

I know QCS isn't a really busy subforum but it might be good to sticky this thread for a while

chances are the Serious Stats Goons are generally also busy doing life

# ? Oct 7, 2019 17:01

Snazzy Frocks: Mar 31, 2003; Scratchmo

this is perhaps the most wholesome QCS thread of all time, great jobs all around!

# ? Oct 7, 2019 22:07

BurgerQuest: Mar 17, 2009; by Jeffrey of YOSPOS

whoa, a sticky! thanks!

# ? Oct 8, 2019 08:10

Harik: Sep 9, 2001; From the hard streets of Moscow
First dog to touch the stars; Plaster Town Cop

Loving the non-uniform distribution of radium autobans.

# ? Oct 8, 2019 21:29

VideoGames: Aug 18, 2003

BurgerQuest posted:

whoa, a sticky! thanks!

It absolutely deserved it. This is a really interesting and cool thread and belongs here. Thank you for making it and coming up with the page to host it

# ? Oct 8, 2019 21:43

Chokes McGee: Aug 7, 2008; This is Urotsuki.

VideoGames posted:

It absolutely deserved it. This is a really interesting and cool thread and belongs here. Thank you for making it and coming up with the page to host it

# ? Oct 9, 2019 06:51

Genesplicer: Oct 19, 2002; I give your invention the worst grade imaginable: An A-minus-minus!; Total Clam

CharlestonJew posted:

I'm surprised genesplicer has so many probes this year. Feels like that dude logs on like, once a month

Nah, I check in almost every day. There was a time when I was just about the only active mod in GBS, so my stats would have been awesome back then. Now we have several mods who are really active, and they tend to beat me to the punch.

# ? Oct 14, 2019 05:37

El_Elegante: Jul 3, 2004; by Jeffrey of YOSPOS; Biscuit Hider

You were an admin at some point, yeah?

# ? Oct 14, 2019 05:57

Genesplicer: Oct 19, 2002; I give your invention the worst grade imaginable: An A-minus-minus!; Total Clam

Nope. Just a humble moderator for General Bullshit.

# ? Oct 15, 2019 03:35

Adbot: ADBOT LOVES YOU

# ? Jun 3, 2024 14:10

b mad at me: Jan 25, 2017

BurgerQuest posted:

Yea ok I had some issues so here are the stats for the 190k+ bans

Radium - to be excluded - 5310 permabans

Top 40 sorted by bans since um, around 2005 according to what's on the lepers colony

I love that Lowtax is at the top. That is fitting and right and good.

# ? Oct 26, 2019 08:35

The Something Awful Forums > Archives > Comedy Goldmine > The Goodmine > stats

«‹›14 »