MangaDex Forums : data & My findings , addendum : General analysis.

[jecroisenlasuperioritedubel]

 

[HoangDinh]

🙈🙉🙉🙉

 

[HoangDinh]

Don't let all the fancy graph(ic)s distract you❗❗❗

What if passlist.txt is a list of our passwords and we just got pwned without even knowing it.

 

[CharaDotArchivist]

Don't let all the fancy graphs distract you❗❗❗

though it would be a total ubel thing to do, passlist.txt is (un)fortunately not a password list

 

[jecroisenlasuperioritedubel]

Now that You did mention It , I may or may not want to get all User data from respective Data-bases to run Anonymized Pass-word data - e.g./ who uses what , or what's the most common , alas ; I won't do It as It's more of an hassle than It's worth , meaning : I won't do such a thing.

 

[Painasbread]

When you click on the whole map - in S.V.G. form you can ctrl+F your username. I'm right below you guys in the picture @jecroisenlasuperioritedubel just posted.

By the way, I'm shocked the great @allenallenallen333 only has two followers. It's probably because he was tling before the following function was implemented. An OG even the forums couldn't accommodate properly.

 

[UnnamedPlayer]

Great work and quite a good idea to do. It's the first time i see social network data science on a manga forum...

Now i'm curious to see how everything correlate to reaction score. Do people with more followers have more reaction score, for example.

 

[jecroisenlasuperioritedubel]

Great work and quite a good idea to do. It's the first time i see social network data science on a manga forum...

Now i'm curious to see how everything correlate to reaction score. Do people with more followers have more reaction score, for example.

Such an intriguing question , surely ; when I re-run another Data Gathering process , I shall implement Reaction scores and other Data fields into the results as wall - thanks for the suggestion.

 

[booterror]

Won't lie, I missed this initially... sorry I'm late. I'm shocked I've made it into any sort of interesting thing.

I'd like to take this moment in the spotlight to thank my fellow follow twin: @Samm0 We've never met but I already think of you as a treasured follow friend. In honor of this, I plan on following you to symbolize our new follow friendship ❤️

To @alimoodyxx thanks for all the likes on my shitposts and congrats on also making this list! I'm proud to reveal to the crowd here that I'm your lone follower.

And finally, thank you @jecroisenlasuperioritedubel for including me in this. I'm glad to have done nothing, contributed only shitposts, and followed you out of solidarity and making me chuckle with this very interesting dive down a rabbit hole... expect a follow.


Lastly my reaction to all this and seeing my name in the list:

 

[alimoodyxx]

honestly i thought i did something wrong but it's not! thankgoodness plus i'm very pretty lurker yes, because my second language english is garbage (and can be incorrect grammar sometimes) and i followed those people because they are everywhere whatever i found manga and manhaw (idk about manhua tbh) and those comments are very very interesting sometimes its funny and information (plus spoilers too, depending on manga i read could be good or bad)

@booterror thanks for following me my man <3

 

[robertelch]

@Angry_Panda
First of all, the scraping is finished and will, as such, put no more strain on the servers.
Secondly, there's no commercial intent. Just a love of data and data analysis.
The only involved person at first was the French eccentric himself, now joined by myself; no other third parties involved.
And as for his mental wellbeing: I don't know anything specific but from my communications with him, he, at least, seems to be in full possession of his mental faculties, even if rather eccentric.
The full dataset will also be published freely once it's completed

 

[robertelch]

@Angry_Panda
First of all, the scraping is finished and will, as such, put no more strain on the servers.
Secondly, there's no commercial intent. Just a love of data and data analysis.
The only involved person at first was the French eccentric himself, now joined by myself; no other third parties involved.
And as for his mental wellbeing: I don't know anything specific but from my communications with him, he, at least, seems to be in full possession of his mental faculties, even if rather eccentric.
The full dataset will also be published freely once it's completed

I'm also ready to answer any other questions to the best of my abilities

 

[Doedoel]

Hope it's readable, @Angry_Panda

 

[robertelch]

Hope it's readable, @Angry_Panda

What I said but more eccentrically

 

[Angry_Panda]

Thanks to both @Doedoel and @robertelch for acting as middlemen.

This may just be (and seems to be) simply an exercise on the part of dubel, which is fine, but as someone who works with marketers, I can say with some authority that no data set is immune from being turned to their purposes if they get their claws into it.

I can also say that based on legal opinions I've read for work, the EU and now some US states really don't care why someone has user data, they expect users to be able to request removal of their data and that those requests will be honored. I would be apprehensive about being in possession of a data set without a clear compliance process in place. YMMV.

 

[Doedoel]

yeah fair 'nuff

 

[sterven]

I'm shocked the great @allenallenallen333 only has two followers.

surprised to see someone mention him in the forums I still remember the disappointment i felt when he dropped iron ladies or did he retire i don't remember which one it was

 

[tilefish]

View attachment 21468After seeing the Visual novel made by @WanderingWanderer , I've decided to do what I do , and indulge in - even if It Some-times get rather problematic , the best : Data acquisition and interpretation , whenceforth ; I may or may not have spent Some time looking around.

RAW DATA :
TOTAL ACCOUNTS VISITED : 1653
FULL FRIENDSHIP LIST : ATTACHED - friends.txt
FULL USERS LIST : ATTACHED - passlist.txt

INTERESTING DATA :

1 653 total accounts in the file.

2 006 directed " who follows whom " links once Every-thing's parsed.

26 % of accounts are total loners - has no followers , follows No-body.

Average account has approx. 1.1 followers and approx. 1.0 followings – so basically Every-one's talking to an Empty room.

8 % of all relationships are reciprocated - i.e./ both parties follow each other.


Wild outliers exist on both ends : mangadex is followed by 367 ( depicted 41 here ) accounts & follows absolutely nobody while alimoodyxx.759213 follows 32 people while only one Sad soul follows back.








Averages :

Metric	Min	25 %	Median	75 %	Max	Mean	Std-Dev
Followers	0	0	0	1	41	1.109	2.487
Following	0	0	0	1	32	1.042	2.307
Degree	0	0	1	2	41	2.152	3.698

Big people ;
@MangaDex : 41 followers ,
@asphyxiav2 : 20 followers ,
@athayanezant : 20 followers ,
@HoangDinh : 19 followers ,
@DavidianMillerian : 19 followers ,
and Serial followers ;
@alimoodyxx : follows 32 , has 1 follower - yea , Professional lurker ,
@booterror & @Samm0 - 20 followings each ,
@tilefish - 18 followings ,
@WanderingWanderer - 17 followings & 10 followers - the best , honestly.

Reciprocity - or lack thereof - is in shambles ; out of 2 006 Directed links , only 161 pairs were Bi-directional , Effectively meaning ; Reciprocity rate is 8 per cent , with translation being : for every 100 follow actions , only 8 are returned , i.e./ It's basically a One-way Shouting contest.

95th Per centile Cut-off : Any user with > 5 followers or followings is already top 5 per cent , in essence ; even after We exclude the freaks , the avg. John MangaDex barely goes above one follower and a - Whooping & Grand - total of 22 people manage more than 2 Genuine 2-way relationships , in other words ; the Social graph is almost flat - It's a ton of Single-edge arenas and an handful of Tiny hubs.


From the chart It can be seen that there's a Deeper - & Secret - MangaDex society.

The 15 most common 3-letter combos : 1. man - 53 2. the - 53 - indeed , a tie 3. ter - 40 4. ing - 37 5. and - 34 6. rea - 34 7. ang - 32 8. ari - 28 9. ste - 26 10. ver - 23 11. der - 23 12. ant - 22 13. tra - 22 14. art - 22 15. ran - 21

Rank​ Suffix​ Raw count​ 1​ san​ 7​ 2​ kun​ 6​ 3​ sama​ 5​ 4-7 ( tie )​ chan , reader , shi , senpai , master​ 3 each​ 8-12 ( tie )​ knight , reaper , gaming , sensei​ 2 each​ 13-n​ every other " wannabe " suffix ( gamer , nii , dono , cat , & c. )​ 1 apiece​

If You're a guest and reading This post before You want to create Your New account , here are Some advices :

• New User-name means You need to choose It well and if You don't want to blend in , avoid any combo Containing man , the , ter , or ing - as They're already living Rent-free in Every-one's names.
• Slapping -san on the end will make You look about as Original as a Naruto Head-band at an Ani-con.
• If You want to be generic , take a name that's exactly around 9 characters long - Arithmetic mean over the Whole Cleaned list , that includes : Under-scores , hyphens , & What-ever Weird Other people love - and make It mantheter or themanter or tertheman or similar , You will be the John MangaDex , ergo ; congratulations , You shall slap a -san to It if It's not enough - theman-san.

Longest name : @theworldaroundusfallseverysinglesecondwespendhere , Shortest name : @3F.

Most gentlemen keep It around 10 characters ; one lunatic stretches to 49 , and a couple of minimalists barely type at all.


Hereinunder is attached the whole map - in S.V.G. form :

and an High-way view of the situation :
​

View attachment 21480


If You are curious , the transition from Inner ring to Outer ring happens by people getting in groups , such as this scene :

Ranging from 4 to 9 people rotating around a Single person - or vice versa.

Data for people are gathered from : Friendship lists - starting from @jecroisenlasuperioritedubel & " Active members " page.

The absolute beauty of these statistics makes me far more giddy than it should.
You get a follow.

 

[sanctuary]

Selling this to the bilderberg group.

 

[jecroisenlasuperioritedubel]

Data-base can be downloaded using This link : https://limewire.com/d/fI3lc#7VyvDHaJgn. Time limit : 1 week.​

RESULTS​

View attachment lang_usage.png English usage is over 80 %.	When People are using emojis , there is not a standard trajectory and seems — mostly — irregular.
A.i. based results show that although sentiment is Mainly based on criticism no Axctive toxicity is observed.	Crying emoji is the Most Popular emoji currently.
Heat-map of Active hours Before & After closing the forums side-by-side for a Calendar year.	Heat-map of Active hours in a given dꜽ.
The Most Active hours in a given dꜽ.	View attachment forecast.pngFore-cast of the activity in MangaDex Forums until 2033.
Rolling & Moving ꜹg. for amount of posts posted Each dꜽ.	Fore-cast closer.
AcF & PAc results ( hourly ).	Inter-arrival-time Distributions ( Log scale ) & entropy by period. Histogram of Inter-arrival-times on log. x-axis.
https://files.catbox.moe/si2eex.png MangaDex Forums Most Used words( 1000 ) 8K Wall-paper.​	https://files.catbox.moe/drmgls.mp4 Cumulative Emoji-usage over time.​