I Made a Graph From My Subscribers to Prove You’re Nerds

If you'd like to support these videos:
/ notdavid
This is definitly the most effort and mathematics I've ever put into calling someone a nerd. I hope you enjoyed it! I also didn't know about this until it was pointed out to me, but youtubeatlas.com has a very similar network to mine. Our networks were made differently but conceptually are similar. Highly recommend you check it out!
Made using #blender as well as Davinci Resolve and Gephi.
DISCLAIMER: My channel is not endorsed by nor associated with any individuals, creators, or organizations featured in this video.
But also I hope that anyone that was included interprates their inclusion as a sign of my appriciation for their content.
Creators included (whose names were not explicitly written on screen):
Jenny Nicholson, Summoning Salt, Matthewmatosis, Technology Connections, Drawfee Show, Super Eyepatch Wolf, Jacob Geller, Hbomberguy, Folding Ideas, Legal Eagle, Polygon, Brian David Gilbert, Karolina Żebrowska, Bernadette Banner.
Content Featured:
The qoute at the end was taken from Matthewmatosis: • Meta Microvideos
The speedruning video was featuring Darbian, but it was from the Summoning Salt channel:
• Mario: The Infamous Hi...
Chapters:
0:00 I need a hobby lol
1:38 Building the network
4:26 Who do you subscribe to?
6:46 co-subscribers and bots
9:29 Proving you're nerds
11:37 The KZbin network
14:20 The KZbin communities
Music (in order):
RPG Store - Chris Doerksen chrisdoerksen.bandcamp.com/al...
Native - HOME open.spotify.com/artist/2exeb...
Killer Vacation - Chris Doerksen (see previous link)
Literally Chill - Florida Skyline floridaskyline.bandcamp.com
Intermission - Stux.io and Vaporwavez stuxio.bandcamp.com/album/qua...
Vague Familiarity - Not David • Vague Familiarity
Atlas - HOME (see previous link)
Power Cycle - Kuraine/ Lena Raine radicaldreamland.bandcamp.com...
Notes:
-Regarding the degree distribution and cutting it off abruptly - in actuality the fact that it is a straight line in log-log is really interesting, but that could be a video on its own so rather than going into it, I just decided to leave it as a joke because this is actually one of the things you would do first in a real network analysis.
-I didn't specify this but when I removed the weakest links from the co-citation network, I also removed any nodes that became disconnected from the network.
-In order to make the youtube network I had to play a bit of a trick. You see, the API does not (as far as I know) let me just randomly pick a user from youtube. Because of this, what I did was to get a random list of my subscribers. Then I get a random list of people that they subscribe to. Then I ask who do these people subscribe to, and that is how I build the network. Essentially I'm adding extra steps to the original method to "get away" from my subscribers. Because of this, it is also possible to get channels that do not subscribe to anyone, leading to an isolated node - these were discarded from the analysis.
- For the community detection the Louvain algorithm returns something like 140 communties. However, many of these are very small and while I didn't do additional testing, I doubt they would be statistically significant. These channels were removed from the animation so that I was left with only something like 8 fairly large communities. They were included in all the other analysis however, they just aren't rendered.
- I've put the co-subscriber network gephi file on github
github.com/notDavidsGit/notda...

[@not_David]

How many channels do you subscribe to?

[@donpantolonez]

57 most of them are specific and not one spammer. I guess I’m letting KZbin to choose the content for me.

[@not_David]

​@@donpantolonez I don't know, I feel like with 57 its more like you are dictating what you want to be recommended (assuming the algorithm cares about subscriptions). Also, users with small number of subscriptions are super interesting to me (from a data point of view) because almost by definition you are unlikely to be subscribed to me, so on my analysis I have very very few people with less than 50 subscriptions. But that doesn't necessarily mean they are rare on youtube as a whole, and its entirely possible they are actually the majority.

[@ajx4]

I am part of the subbed to 1000+ channels people. I am 21, my account was made around 2012 ish during grade school, my subscription list is a collection of everything I have been interested in most of my life, and I have def unsubbed from things I disassociated with (got canceled and I didn't stay or changed their main content focus to something I don't care about) It's interesting to see you ere on the side of "they are bot accounts" because the world I live in is mostly Yotube, and I think most people I interact with (gamers/nerds) are like this as well I also don't understand NOT subbing to channels that make good (or even just A good) video. Like I subbed to you off of just this one random recommendation, because I know I'm going to be interested in your channel going forward. Sub list is somewhere I can go back and say 'hey that's where I learned that, let me go back and see what else they have made recently. do I interact with the content of all of them? no not necessarily, but it's like a bookmark folder for me, then I can take a sec and figure out what/why and maybe catch-up on recent vids or whtvr. Hope that helps lol. Thank you for the vid. (Also what's up with the dig at drawfee what did they do?)

[@THE_ONLY_REAL_WAFFLE]

I'm part of the 1000+ channels subscribed to

[@donpantolonez]

@@not_David I’ve mechanical engineering back ground and using KZbin like a university more than 10 years. Your content fits the bill for me (fast, technical) and quality over quantity. Algorithm definitely cares about the subscriptions but also driven by recently consumed content in my opinion. When my coffee machine broke I got coffee machine videos recommended for a month. I don’t use google as a search engine so KZbin couldn’t get any data from there. Most of my subs are when channels have between 5 to 100 k subscribers and because of your video quality I think your trajectory is up and up, enjoy the ride.

[@gvarph7212]

If you had to use network theory to realize you too are a nerd, you definitely are a nerd

[@not_David]

I'm still in denial, likely a statistical fluctuation.

[@Mittzys]

​@@not_David"Maybe that also hints that I, too, could potentially be a nerd?" I believe that you saying the above quote instead of "looks like I'm a nerd because you guys are too" Is also evidence that you are a nerd

[@knpark2025]

@@not_DavidAs a nerd who is subscribed to all top 10 channels you are one of us.

[@gianlaager1662]

@@not_Davidthe fun thing any one who calculated that significance of that found that they were a nerd. Does that mean that every body is a nerd or is that just selection bias? I guess we’ll never know.

[@elsaarcilla5664]

You could try what I did and pretend to not be a nerd. But it doesn't work lol @@not_David

[@CarterPatterson1228]

As a dishwashing powder user, "through the magic of rendering two of them" killed me lmao

[@watson-disambiguation]

Are you using too much dishwashing powder? :p

[@chasekubesh9241]

IM SO GLAD THIS WAS THE TOP COMMENT, I JUST DIED FROM THAT AND FEEL SO VALIDATED NOW LMAO

[@haph2087]

Same!

[@SciNonFi]

Good to see other people not skipping a step

[@human.earthling]

And why is your car using red brake light turn as turn signals? Stop it!

[@acenio654]

People generally tend not to unsubscribe and rather just gradually watch the channel way less over time. This paired with the fact that some people have had their KZbin accounts for a *long* time would result in a lot of subscriptions

[@blairdactyl]

yeah, I'm subscribed to people I don't even remember watching, but I just don't get them recommended anymore so I don't notice :P

[@PFnove]

i am subscribed to at least a few thousand channels (on this account, i also have another account that i've been using since around 2016) and i only watch whatever's recommended, i subscribe to support the creators and not to see more of their content (there are definitely a few exceptions) i now have tens of tabs open of channels i want to watch later

[@Commenter339]

Not David: _spends hours upon hours doing a ton of data analysis, research, rendering, visual representation, writing and editing and forms it all into a visually and audially pleasing very long video_ also Not David: ... I think I might be a nerd 😳

[@cancername]

To quote a certain German bread: Nerd party!

[@davidelodovichetti4738]

When you called me a bot, at first i felt offended. Then you described the problems that bots caused, and i realised that i am, in fact, a bot to most videos i watch

[@tisdali]

So this is your attempt to not be a bot?

[@ninjakiwigames5418]

Well... Actually even you watching videos is better than a bot. A bot just subscribes and gives you 0 views, 0 watch time. You don't have to always like and comment (although those help) to support the creator, as watch time is the most important thing.

[@TheUltraDavDav]

i've spent about 6+ years without commenting i was more bot than ever, feels refreshing really. not sure why i stopped. but im glad to be back, also makes for a more memorable watching experience. not just everything is television.

[@Bigleyp]

@@tisdaliyeah

[@TheOne-jg7fd]

@@TheUltraDavDav" not just everything is television"......im stealing this gpd sir. its mine now but ill give you the credits and tell people where i got it from. it simple but it says a lot

[@GreatNini]

Wait - I wrongly assumed that my taste in KZbin videos is kind of unique, because I know no other person in real life who shares my wide spreaded interest of nerdy videos, and in this video I must learn that I am not only not unique but am in fact probably watching the same videos with the same people all the time but we don't know about each other?? Insane! Glad you broadened my perspective, I feel less alone now. And site note to being subscribed to many channels: I have "collected" channels I like over the years from changing interests but never bothered to unsubscribe to people I am not actively following anymore. Keep up the great content

[@not_David]

love this comment :) One thing I would love to see is if there was a 'sort by when you subscribed' thing and you could go back and see how tastes/interests evolved over time.

[@GreatNini]

@not_David Oh that would be super interesting! Maybe you could use an API- oh wait :D thanks for this awesome video, I am always blown away by your visualization skills and your general asthetic. The color paletes, the fonts, the shadows, everythign moves so smoothly... really really impressive

[@veryrealpersonwhoisreal]

I've noticed a lot that when a person has seen one video or channel I mentioned they have also seen lots of other channels and videos I have seen.

[@SkyfishArt]

you should try visiting a makerspace, i couldn’t avoid this demographic if i tried.

[@GoingtoHecq]

So if the algorithm is showing lots of stuff to lots of people at about the same time, then it really is orchestrating our viewership. It is a literally capitalist machine making us watch as many ads as it can to get revenue. It feels really bad. It is literally a manipulation machine. It's not great. Even worse is how many years of experience they already have in making and designing algorithms to manipulate people en masse.

[@hoodie_cat]

Guess I'm a bot

[@isle_of_violets]

same

[@Martcapt]

Beep boop

[@octocube3607]

Imagine meticulously organizing all the channels you are subscribed to. I just subscribe to random channel I found cool, never unsubscribe

[@Yellowsam4145]

Sigh, I guess I have to clean out my subs, but this channel is staying.

[@lol_cerdef]

I wonder how many people you can subscribe to

[@queenofthesalt5199]

Honestly terrified me when pretty much the first 10-15 channels you mentioned were all some of my favorite stuff of all time. Also, finding smaller channels in there that you didn’t mention, but included in the graphics, like DeepBlueInk, felt awesome. Absolutely subscribing.

[@iamthinking2252_]

Great Scott!

[@Chaotic-Toast]

Yeah, i did not expect to see DeepBlueInk and SSS in there but there they are lol

[@EKUL34]

As someone who subscribes to 912 channels (now 913) and uses powder to wash items, I can assure you I am not a bot. But this is the first time watching a video of yours, so congrats on making it into my very exclusive network.

[@PeterBarnes2]

I thought I was special with my 881 channels, when I heard 10% subscribed to over 1000 I got disappointed. At least I know I'm not _that_ chronically online, now. I'd've liked to've seen more on the outliers, like what channels with the least total subscriptions had the most Not_David subscribers, and what channels had the most total subscriptions relative to Not_David subscribers.

[@xgozulx]

same xD, I've been in youtube for 16 years though, so that is just a couple subscriptions a month

[@Kehelekarakh]

I'm currently at 1012 and I have absolutely no idea how I ended up here. I do wonder exactly what commonalities exist within the subset of people who are at 1000+ subs (besides potentially being bots)

[@l0kk016]

​@@Kehelekarakh here's a few that come to mind: 1 - actually subscribing to channels - a lot of people have been around the internet long enough to have enjoy way more channels than that, but I believe a minority actually subscribes to these channels. Most people just use the 'recommended" algorithm to go through their phases, and eventually when they move on there's no trace of having subbed for these channels. 2 - varied interests - self explanatory, there's definitely thousands of channels per niche but I bet most people who have a lot of channels dip into multiple communities. 3 - more acceptable/ supportive of small channels - as KZbin is VERY top heavy, if you're subscribed to 10k-40k sub channels you're probably already a minority, much less a large amount of such channels. 4 - not using the "subscribed" tab as much as most people - a bit of a guess, but I think a lot of users are "conservatives" when it comes to subscriptions, as to keep their own bubbles of easily accessible contents... easily accessible. I would believe that this effect is getting less relevant as time goes on and everyone has years and years of internet experience accumulated, common-folk gradually sub to more channels and "terminally-online" folk get used to navigating the chaotic randomness...

[@Pyxyty]

As someone with around 70 ish subbed accts, mainly because i unsubscribe to the channels i stop keeping up with, i genuinely am baffled with how you guys don't get annoyed with the spam in your sub boxes. Im really curious, how do you guys easily find what videos you wanna watch? Do you just let the algorithm serve up what it thinks you'll like or do you still fish through your sub box even if there are channel videos there that you dont watch anymore? Super curious :D

[@knpark2025]

I am a grad student who used to crunch numbers with millions of patent-to-patent citations. This kind of network looks like what I was asked to ultimately work for. With the result you've made with just 2,000 of us, I can finally have a closure about it and be confident to say that my advisor was an absolute asshole for asking me to do something like this with hundreds of thousands of patent documentations.

[@not_David]

I genuinely feel for you wow

[@chloesibilla8199]

Y'all seriously need a union

[@SebastianLopez-nh1rr]

You have my sympathies

[@fa-pm5dr]

i learnt freaking c++ to be able to crunch these kind of network algorithms, specifically calculating distributions of eccentricities. O(n^2) too. Several of them took more than a day. Jesus christ.

[@kellychuba]

this is the bad place!!

[@Mezza_Luca]

"Through the magic of rendering two of them" Okay you got me, I laughed way too hard at that. I've absolutely binged watched that snarky toaster guy.

[@bagelgeuse5736]

5:55 My educated guess to account for channel size would be to compare the probability of a random user being subbed to a given channel vs your audience's probability of being subbed to that same channel.

[@Anonymus_celebrity]

I would really like to see a graph with that method

[@Eutrofication]

Im so floored by how when you reference something that feels weird and specific. It feels unbearably niche in real life, but we really are all weird youtube addicts And id love to add, all the things i learnt about networks is so sick

[@chopczyk374]

Hi, my name is @chopczyk374 and I have over 600 subscriptions. I'am also a youtube addict.

[@andreas4010]

By the magic of making 2 of them Lol

[@hleth4888]

I don't like it, the system knows to much. I'm gonna have to go dark

[@justinpatterson5291]

He speaks r/oddlyspecific...

[@MultiArtStyleTWT]

I felt the same way too, but if we do the *mathzing* then yt has 122 million daily users, so the odds are that a good chunk will be into the same niche channels

[@likebot.]

"You are not representative of the wider KZbin audience" is probably the best compliment I've ever received - and the most reassuring! Thanks, bru. BTW, has this project resulted in you finding a channel or two to add to your subscriptions? I bet there were a few temptations out there.

[@not_David]

oh thats a great question... I dont think there were any new channels that I "found" because most were just names to me, so I didn't really notice them unless I saw it was a funny name but that didn't really mean I wanted to watch them. There were channels that I knew of and then started finally watching because they were showing up so often. I think the main example would be acollierastro. If anything tho what it really made me do is go back and watch old channels I had kinda forgotten about until I saw them in the network.

[@likebot.]

@@theaxer3751 I've heard of Mr. Beast. He was in a Mark Rober video.

[@haiperbus]

​@@theaxer3751yeah, opening YT on a vrowser I haven't signed in on is a terrifying barrage of nonsence

[@caam0000]

​@@likebot.Sure it wasn't Mr. Beat? Or Mr. Bean?

[@likebot.]

@@caam0000 Were either of them working on the Team Trees project? I think that's where I heard of him.

[@killuazoldyckhunter]

I loved the video and I'm dying for the 4 letter word 😭

[@kevincsellak296]

I'm one of those people who are subscribed to 1000 channels. It's because I want to be able to find a channel again if I think of a particular video again after months or years have passed, and because a lot of channels only upload very sparingly (and I'd forget their names otherwise), and because there are a bunch of channels whose genre of content I don't consistently like, but that I do like visiting when I'm in the right mood once every few months. Scootertrix Studios, Nemean, and Biblaridion are examples of these three types.

[@robertridley-fj8zz]

Or, you add it to a favorites list?

[@eccentricity23]

Everyone else has successfully conveyed how lovely the topic of this video is, but you deserve an insane amount of appreciation for how visually stunning this is. I especially liked the blur effect to focus on KZbin communities in the last portion. I hope this channel gets a lot more attention in the future - you certainly deserve it!

[@not_David]

thank you! I was genuinely worried about that animation in particular because I was worried it would be difficult to parse for colour-blind individuals. Eventually I figured out how to also outline the particular communities so I hope that helped...

[@3nertia]

@@not_David It would be lovely if we could poll your subscribers and see which ones have which versions of colorblindness :D

[@emma_loomis]

​@not_David As a colorblind person (with a weird colorblindness) I think it's super cool when people think about us. The world was not built for us haha also I would 100 % participate in that poll ​@@3nertia

[@zergosssss]

to even think of this while making your video surely proves you are a great person@@not_David

[@garrettmillard525]

What do you use for this? @@not_David

[@Lotsiron]

I felt so targeted it's scary in an awesome way. Popular science channels was expected but the more niche stuff I've been watching being exposed felt like watching a magic trick. Like a magician asking "is this your card" you are asking me if these are my interests.

[@TheUltraDavDav]

fr tho. that 9:49 if you stop and really analyze, man can you recognize TONS of names. i wish he published that list in specific so that i could see the channels i will probaby most likely find out later. fells like skiping some steps and getting to the juice a lot more quicker.

[@yeeterguy9295]

"Do you want to enable personalized ad?"

[@567secret]

15:25 Is is driving anyone else up the wall that he missed green?

[@not_David]

me

[@carafurry7862]

15:09 James is the name of the odd ones out. He is our animation overlord. Without him you don't have story telling, or wacky entertaining animations.

[@E_D___]

Yeah, but he looked at usernames - So it means that a lot of people who are in that community decided to use his name in their username (?)

[@KubickQ]

all the jameses united

[@tno1990]

i thought about James Baxter

[@LemonbreadSC]

@@E_D___ When i was a big odd1sout fan at like 11-12 years old (almost a decade ago) I had the name "JamesOdd1Fan" or something along those lines. It's not impossible.

[@stemwaffle]

I think this is the nicest way anyone's ever called me a nerd in my life. On an interesting note, I find it hilarious that out of such a diverse sample size, that is pulled rom a huge area, James is common enough used word in usernames FOR A COMPUTER TO PICK IT OUT?!?! lol

[@WoolyCow]

if i had to guess why james, itd probably just be theodd1sout stans :>

[@MultiArtStyleTWT]

@@WoolyCow yeah thats exactly why, "Animation" was also in that community not to mention his podcast also in that community

[@conando025]

It's so fun watching this video and being like 'oh hey I know that channel'

[@not_David]

this was my favourite part about making the video as well, no matter how niche the channel I could think of I could always find someone else in the network that watched it which was ... strangely comforting

[@siliconsulfide8]

Even more fun if the channels have pretty much nothing to do with eachother (and if you're not part the Not David network). "Oh, look, that's Scishow! Yay! NileRed, yeah, might've guessed." "Cool, it's Tom Scott." "LegalEagle? Oh, I wanted to watch one of their/his latest videos!" "TierZoo??? Oh, look, I'm still subbed to them!" "Karolina Żebrowska?? Cool." " _Jaiden?!_ " ...maybe the reccomendations I'm getting from KZbin aren't all so random after all

[@leave-a-comment-at-the-door]

I was basically the inverse, I recognized so many of the channels shown on screen so I was like 'oh wait what's that channel' when there was a icon I didn't recognize.

[@justinwatson1510]

"... *I* might be a nerd" he said, looking up from the math homework he asssigned himself.

[@death_of_cosmos]

This just popped into my recommended and I haven’t felt so called out before. And as someone who is subscribed to practically a thousand channels many are smaller newer channels with little content at the time but the majority are a part of mass subscriptions when going through some content faze. I can say that whenever I do revisit these channels I truly enjoy them.

[@hvok99]

No matter how 'connected' we are as users, the vast majority of people experience the content alone. This video reminded me how connected we all are 💚

[@onedeadsaint]

my immediate reaction to your comment was to chuckle. i will now be _not_ thinking about how sad that sounds.

[@monad_tcp]

I wish there was a convention on of the channels in this specific network, but then probably I would have to move to United Kingdom or something

[@richtigmann1]

my god you're right

[@Lishtenbird]

I feel a bigger disclaimer was worth on the "those with public profiles" part: people who chose to guard their privacy (and were aware of the option in the first place), as a group, likely have different (say, more tech-savvy and less mainstream) preferences. But those tendencies won't get represented in the sample - because, well, data from them just wasn't available.

[@Providence83]

"Less mainstream" lol I was an alcoholic, eating balls of mozzarella on a mattress _on the floor,_ watching videos of people *dying!*

[@liam8370]

and that's not all I use two KZbin accounts feed the algorithm to tailor it for what I want and use incognito for everything I don't want in my feed like I watch a little too many tech videos and almost always learning some tools plus my tendencies to randomly get curious about something and spend the day on it so it would be hard for even KZbin to find out what I'm like cuz I'm only showing them what I want them to know about me and that is I'm a science nerd above everything who likes anime my second identify is a gamer who likes to use songs for background music while coding as for who I really am "just an average guy who claims to love math but is a philosopher who's obsessed with death"

[@dimaryk11]

Kinda like survivorship bias

[@MultiArtStyleTWT]

@@liam8370 yeah that when you know your like a *NERD* nerd, you use incognito not for rule 34 ow2, but for *tech videos*

[@leave-a-comment-at-the-door]

@@liam8370 same sort of thing I do; have different accounts for the kinds of videos I want to watch in different moods; so that I can have one account for learning things; one for brain-off spammy stuff; one for each of various communities; and so on and so forth.

[@Luredreier]

9:12 I'm not a bot, yet I'm running up against the upper limit of the number of subscriptions that KZbin allows...

[@aryanmahajan9851]

Ya how

[@tweak3871]

Perspective on some of the analysis from a ML professional who used to specialize in recommendations systems: - It takes a wild amount of skill to communicate and present information as complicated as this is, and I commend you for it. Love this stuff, please keep making more. - On the note of "compatibility", I don't know what analytical tools exist for comparing pairs of ego networks (not a network theory guy), but any notion of channel similarity should probably include directly comparing channels. By measuring just using overlap, as you pointed out, we become very susceptible to popularity bias. Therefore I would propose developing a notion of similarity by figuring out a channel's "average viewer". A somewhat fast and easy way to do this would be to randomly sample a few thousand subscribers from each channel you want to compare yourself to, one-hot-encode who they subscribe to, calculate an "average subscriber vector" by summing up and dividing by total magnitude. You can then compare channels via cosine similarity between their average subscriber vectors to get a metric of similarity. Channels that are more susceptible to popularity bias therefore will be more different from your channel as there should be a broader set of channels that their subscribers subscribe to on average, whereas more niche channels that have a lot of overlap with yours will have higher similarity. I imagine there is some network theory equivalent to what I just described, but that's pretty close to how I would measure similarity amongst users/items in an interaction matrix. Here we just don't have the benefit of having the entire user video interactions at our disposal, so we have to sample. - On the note of "see how modular youtube is when I apply network theory to it?", you can probably already guess the criticism which is, if you apply a technique explicitly designed to make the borders more obvious and clean, and to ignore "noisier" connections, how do we know it's actually that modular and not the methodology engineering it? I don't know the method you described, but just generally in my experience, unsupervised methods make algorithmic choices that have this way of making people think patterns that are there are much stronger than they are. But I also do believe that in this circumstance, there is likely some truth being communicated about the data there, even if it isn't likely the whole truth, so I don't want to make it seem like it doesn't have any merit, I'm just a bit allergic to strong assertions when unsupervised algorithms are involved. Great video! Earned a sub from me :)

[@not_David]

Thanks for the all the cool comments! Regarding the last one in particular, its a great point and yeah I had this thought too. The louvain algorithm will give non-overlapping communities no matter what. There do exist community detection methods that do not do this, but they are beyond this video. However, other (non-community related) metrics also seem to support modularity as well as other independent studies (e.g., the youtubeatlas which I wish i had known about prior to making this video). So while I agree this analysis would be pretty shaky evidence for youtube being modular, other evidence points towards this as well. So I think if you had to place youtube on a spectrum from non-modular to modular, I would put money on it leaning towards the latter.

[@tweak3871]

Interesting! I'll look into that! Again keep up the great work mate@@not_David

[@scifii_official]

Maybe "James" in the "animation and podcasting" community is in reference to the creator TheOddOnesOut, pointing to that community consisting of animated story-time creators. Great video! I've been thinking about network theory in the abstract for a while but didn't have a name to put to it. Definitely will look into it more now.

[@not_David]

I was thinking the same, I just never realized he was that big. Maybe I was expecting Jaiden if anything else. Network theory is absolutely amazing. I find it relatively intuative and the math is really simple. I HIGHLY recommend Mark Newmans introductory book to it. The first fairly section of it is a broad, math-free over-view of where and how networks are studied (e.g., social networks, the internet, electricity grids). Its also called Graph theory by mathematicians, but those books tend to be much more mathematically dense even though its more or less the same thing.

[@caspergotlost]

James is widely considered the OG animation story time channel, and has a significantly higher subscriber count, but I think Jaidens more nerdy stories fit in with this community's vibe better.

[@scifii_official]

@@not_David Thank you for the recommendations! I've always been interested in math (I watched ViHart and Numberphile growing up) and I'm in university for my Computer Science degree, so I'll be sure to check out both the intuitive and math-centered resources.

[@Arnaz87]

@@caspergotlost that is such an interesting conclussion and I love this channel for making it possible in the comment of its videos, along with all the other interesting discussion going on

[@davidanoble]

6 Degrees of Jaiden Animations. I think James is bigger within the local area, but I'd bet Jaiden has more connections outside. I first heard of Jaiden from watching Ryukahr, who is a gamer and mostly gets content from Super Mario Maker and its sequel. Which seems odd.

[@Mr_Soleo]

I'm over here wracking my brain about what the most common thing is, and then I get slapped in the face by the realization that it's almost certainly ASMR. Amazing video, beautiful visualizations, If it is ASMR I hate it here.

[@not_David]

you guessed it haha (unfortunently)

[@mehnemjeff4793]

@@not_David oh come on, I was really hoping it was game since 2 of the identified communities had some form of gaming in their names.

[@user-rl4rl7sv2y]

@@not_David oh no...

[@catcat2607]

@@mehnemjeff4793gaming is for fun. Getting put to sleep is crucial for us insomniacs

[@spiralspark8523]

​@@not_DavidASMR is from hell and I am glad you agree 😂

[@FirstNameLastName-gh9iw]

15:18 ah yes my favorite community James(?)

[@durdleduc8520]

i have never, ever in my youtube-viewing career been more called out by any singular video. from GMTK to Summoning Salt to fucking Drawfee again, to getting called out for the fact that I never clean out my subscribed channels.... i am so mad. thank you. (i'm at 223. i really need to go through them again.)

[@not_David]

my favourite comment so far

[@suncat530]

i'm probably one of the people with 1000+ subscriptions, 'cos i never clean up my subscription list... the last time i tried to export it it froze my app, lol i just like to sub to people, but because my sub list is unmanagebly large, i have to rely on algorithm on home page to give me videos from people i want to watch.

[@Omninfinity]

Me and my 2k subscription list (I can't add anymore unless I have more subs so I just bookmark channels I want to subcribe)

[@Junniebug]

​@@suncat530sameeee

[@MultiArtStyleTWT]

good luck! im at 550+ , so its too late for me!

[@pielover267]

Oh no, I'm gonna be bothered by never finding out who that one very weird channel is or what the most common word was for a long time lol. I was expecting the word to be revealed in the description, or comments or end of the video or something. My guess is "tube" being used as a suffix, rather than a standalone word, or maybe "Vevo" but I don't know if that's a word or an acronym

[@not_David]

I'll give you a hint, it starts with A and ends with R

[@monopan2606]

@@not_David I fear I'd hear a "bing" sound if one were to buy an "M". Thank you for the hint to satisfy my need for an answer. On to finding out about the blurred channel next, I'm afraid...

[@Max24871]

This comment needs more interaction, I had to scroll way too long to find it

[@leave-a-comment-at-the-door]

someone else found it was asmr

[@bob_mcbobberson]

@@leave-a-comment-at-the-door finally

[@johannguentherprzewalski]

I am very grateful to have found your channel. Seeing your animations, the way you explain mathematical topics in an approachable and authentic way is probably the perfect example of the things I want to achieve myself. I am looking forward to your future work!

[@AlmondJoy32]

As someone who personally related to every channel you shouted out as your own personal favorite, it's pretty amazing to see the youtube algorithm lead me to this channel too! Really good vid, taught me a lot about netowork theory and this community!

[@nathanlaleff4273]

The reason people are subbed to 1000+ channels is actually pretty simple. Many creators make content you might like, but at the same time, you might not like ALL of that creators' content. That logically means that you might lean towards having a large collection bin of which you'll filter out the content that interests you. It's like gaming or film. There are a lot of genres, and you might not like all of that genre's catalog, so you browse through your preferred genre until you find what you want. At which point you'll binge for the next week before starting again.

[@gtbkts]

Exactly

[@OoMikkioO]

Good point. Just because I enjoyed RDJ in Oppenheimer doesn't mean I wanna watch Doctor Dolittle just because he's in it. Kinda same thought process

[@noomade]

Hi fellow Nerds! ❤

[@alyosharomanov]

hi

[@RagaarAshnod]

I am in this comment and I appreciate it.

[@amazingbutno5303]

Hey!

[@not_David]

I guess I am here too :(

[@noomade]

@@not_David good place to be!

[@EPMTUNES]

Hey not David. I’m subscribed to 3k+ channels and a real person. I love this video and the videos you’ve made. It’s great being able to quantify these links while also talking in general about graphs

[@OrangeManta104]

3k+, HOW???? Im subscribed to 21 and three or four of them are second channels, so I’m subscribed to about 17-18 accounts. How do you even find 3000+ channels that interest you? Btw. No hate

[@sebasgovel]

This is the quality content i watch ads for (keep up the blender visuals!) I'd love a video combining the process of doing the visuals and the math behind it!

[@physicsforthebirds]

16:16 Bird. Bird is the 4-letter word, it must be.

[@not_David]

got it in one!

[@jessehammer123]

Why? Why is “bird” the word? I can’t believe that “bird” is the word! I had no idea, but it seems like, apart from myself, everybody knows that “bird” is the word. But seriously, though- why is it “bird”?

[@throstlewanion]

Never in a million years would I have guessed that Drawfee would’ve been featured in this video

[@not_David]

just trying to contribue to that 2million grind

[@tinkersdinkers]

LMAO ME TOO guess we're all little weird podcast went into science but still had an artist phase when we were younger and still enjoy that creativity 😋

[@tennicksalvarez9079]

Or even shawn!

[@tricksterdim]

Awesome video! The blender graphics compliment the video very well too, they are very smooth. Well done!

[@jeffmorrison5834]

Cool video man. Your explanations of group detection made the process way more intuitive than other formal definitions ive came across. Congrats on being my 1,467th subbed channel!!

[@not_David]

[QUESTION] The part about the 1000+ subscribers still really bothers me. If you're watching right now, how many people (roughly) do you subscribe to? (ok I didn't expect that many replies lol, I can't respond to all of them but they are interesting numbers coming in thank you!) Also - please let me know if the audio has improved (or especially if it has not). I spent a lot of time trying to make sure it was better but I'm still not sure.

[@immysoun]

990

[@creepyshadow55]

85

[@eta.tauri32]

I'm at 860.

[@Pyxyty]

Not much, around 70-ish. Lots of creators that upload every 3-6 months with polished videos, then some esports channels, some vloggers, some animators, some gamers. I make sure to keep it clean and remove any that I lose interest in.

[@DogsockDogsock-ll7ed]

14

[@brycewashington5022]

Honestly top tier KZbinr in the making. Amazing content and production. I’m currently an undergrad studying CS and you’ve just inspired me to take a class on network theory. Keep up the great work!

[@not_David]

yes! best compliment I can think of. Its genuinely such a fun/interesting field, I'm bummed I had never taken it in my undergrad.

[@fightocondria]

Highly agree, you're punching well above your subscriber weight. Keep this up and you'll be at 1m in no time

[@fightocondria]

I wish I was still in school. Data science is huge right now, and insanely interesting

[@joshuaadesida]

Just found this KZbin page, but the accuracy is amazing 😂, officially subscribed

[@beesinpyjamas9617]

Discovering your channel for the first time and yet this is an exactly accurate representation of the channels I watch, I guess the algorithm knows exactly what kind of person I am, along with the swarm of all the other people who are saying the same thing who are also first timers to this channel

[@caspergotlost]

As someone who has kept a few pages in my notebook where I write channels I watch and mark the day I started watching them, and connect them with lines every time there's a crossover, this video is amazing. It also gave me a crisis trying to decide whether to add all the mentioned connections (I didnt) I have given up on dating things, it got OOC, but I still update the network, and it's interesting to say the least. it reminds you that algorithms work in mysterious ways, and also that just because a creator posts about one factor of their interests, they're a human with many interests, and collaborators show that off, bc people you don't expect to have heard of each other can be amazing friends

[@not_David]

really well said

[@HelPfeffer]

15:13 I think James refers to TheOdd1sOut

[@florianschneider3982]

right group but why would his name appear?

[@DVnc_]

He is that popular, lol​@@florianschneider3982

[@nick_9]

You deserve Millions of Subs! This is the first video that I see from you and I'm genuinely shocked about the production quality. It's so visually pleasing while teaching us a bit of Network theory as a treat. I love it!

[@not_David]

thank you for the kind words :) (and on the uhmms video as well so I'm not spamming you lol)

[@4bSix86f61]

4:52 "Powder based detergents" killed me 🤣

[@jeffersonye7572]

I get such whiplash when I see some of my friends’ KZbin home pages. There’s so many different, separate communities on this website that we’re not even aware of because we’re walled off by what the KZbin algorithm knows we like to watch the most.

[@not_David]

yes! When I was telling my friend about this video in its early stages we shared our subscription feeds and I was like "i've never heard of any of these" and they are like 10M sub channels. Its kind of mindblowing to me.

[@fallow64]

This is an amazing video. A broad look into network theory, making many enjoyable references, to expanding the range to the human brain, and connecting your own personal life from a question as simple as “What does my audience look like?”

[@not_David]

very nice comment, thank you :)

[@joeybrown9952]

first video i’ve seen of yours, absolutely incredible, youtube knows what it’s doing by recommending your channel to me, the network looked extremely similar to my youtube feed. i turned off my history years ago and have a private account, but it still found a way to connect me to you. this vid warranted an instant subscribe

[@raisedincalifornia1828]

This is the first video of yours that has even ever come across my feed. So glad I came across it.

[@throughcolouredglasses9300]

I have never seen this channel before and don't follow much computer algorithm stuff, so I was intrigued to see why this was the first video on my starting page. Turns out according to your formula I should absolutely be the target group, seeing how I love Jacob Geller, Bryan David Gilbert, SciShow, Tom Scott, Drawfee, Bernadette Banner and many more of the channels you showed. I guess I will listen to the power of the Algorithm and check out another of your videos :)

[@MaidenOfAir]

Bernadette was probably the *most* surprising mention in the video. I wasn't subscribed to not david before this video, and that mention hit me like a truck

[@Nefi424]

Seeing the tiny Drawfee bubble sandwiched between the titanic Veritasium and Tom Scott ones was really funny to me, for some reason

[@illuminum8576]

This editing style is amazing!

[@jonathanettinger3104]

i'm subbed to more than 2k channels, but your vids are truely special and i really enjoy them

[@not_David]

very kind words, thank you :)

[@IMPERIALYT]

Super interesting - I'm usually not a data guy but you made this very digestible

[@not_David]

my favourite type of compliment, thank you

[@fightocondria]

Sounds like you didn't know you were a data guy. ;)

[@not_David]

^ this

[@tnijoo5109]

I have pretty severe ADHD and it makes me feel a lot better about myself knowing there are plenty of other people subscribed to over 1,000 channels, assuming they’re people and not bots. Thank you. ❤

[@wang78739]

Hey, another ADHD-er with over 1,000 channels here! But yeah, when you go down those hyperfocused interest "rabbit holes", you suddenly find yourself subscribing to some pretty niche channels most people don't even know exist (like I've got a whole subset of Japanese vending machine people, sheep farmer channels, historical fashion/garment creation youtubers, motorcycle maintenance, fly tying for fishing, cold process soap makers etc). Have always been meaning to do a purge but... eh. I sort of see it as my way of bookmarking channels that I wouldn't be able to find again with the search bar/recommendations these days. The thing is that because they tend to be pretty niche channels, they don't actually upload too often (or are dead channels that haven't uploaded in years), so my subscription box doesn't get flooded. Plus I don't watch every single video that hits it, only the stuff that catches my eye that day. I do have the bell on for Not David though, since they are always worth the watch! :D

[@prageruwu7148]

Same here

[@ItsAsparageese]

Same all around! Also, in case anyone was curious, the limit for your Watch Later is 5000 videos ... Guess how I know 😂 I assume it applies to other playlists too, but I haven't hit it on any others yet. I'm sure I will at some point though lol

[@sdestroyer6135]

Hell yeah

[@sdestroyer6135]

@@ItsAsparageeseit does 😢, on a more serious note I have around 40ish playlists with all my interests and just rotate around

[@Leander_]

This was fantastically well presented, awesome video! I think you went in to just enough (mathematical) depth for the network analysis to keep things engaging and concise. You definitely gained a new sub :)

[@ducktor6160]

Amazing video! I had a subject on my discourse master's degree with Scolari, developer of the interface theory, which works essentially like the network theory, and it's an eye opener for sure! Hope your channel keeps on growing 💗

[@bromanned7069]

I love that somewhere on KZbin there is a Barbeque Gaming community

[@insanelyinshape]

I'll be honest and say I used to collect KZbin channels like they were candy. I had easily 1300 channels I was subbed to, while watching only a fraction of those on any given day. However, I made a major purge several years ago and since then I've been much more selective about who I'm subscribed to. Glad to have seen this video and know that I'm in good company!

[@MINECRAFTLOVER4000]

im currently subbed to close to 5k channels lol

[@chubs2312]

I purge regularly. If I scroll past 2 videos from a channel on my sub page I unsubscribe. I end up with channels That only post once a month or week. But at the same time I still watch the bigger channels that post more often eg MBKHB

[@jtessler8956]

Just under 8k subscriptions for me

[@CullenCraft]

This was the first video I've seen of yours... And shockingly I'm subscribed to nearly ALL of the channels you listed by name. Except the Victorian fashion ones - maybe that's my next adventure

[@DonJarvius]

This is an incredible video. I’m don’t know how I was recommended this, but now I’m so glad I was. Subscribed!

[@cirecrux]

The real network was the friends we made along the way

[@HelPfeffer]

HAHA Yes

[@andeccan]

That's true

[@WangleLine]

aww :3

[@evanbarnes9984]

Your animations are incredible! I wonder what it would take to make this same project happen on a much larger scale. I'd love a nerd-focused co-citation network to replace KZbin's broken recommendation algorithm. It would be be really cool if we could just have an open source version of the recommendation algorithm that actually only focused on connecting you with creators that you might like, rather than pushing advertisements and paid placements. I would also love to be able to ask a recommendation algorithm to send me to a region of KZbin that I'd never experienced before and wouldn't normally find through a typical recommendation network.

[@not_David]

I'm not entirely sure. I think its too easy to over-do it. There was a whole cut section on how actually what i really appriciated where the "bridge" channels that couldn't really be classified into one community. For example, the channel Sebastian Lague is very much on the border of computer science, art, and even life science, and I think we need more of those things to actually get us out of our "community" edit: to be clear, i do think the algorithm could use improvement, but it is a monumental task and i think theres a lot of nuances that need to be taken into account

[@chalkchalkson5639]

@@not_DavidDo you have the data available somewhere? I wonder what channels are involved in inter community connections

[@rahipatel812]

This was extremely well done and taught me something Ive never thought about before! Does anyone know what the 4 letters were? Did I miss something?

[@Bumbledick]

I was thinking the same thing the only thing that comes to mind is VEVO

[@jasonglasser1621]

2 of my favorite channels had big enough circles on your map to be visible, *and they're the same guy* (Wendover and HAI). Plus a massive Tom Scott circle? Yep, looks like I ought to be subscribed to you, your people are my people.

[@tylerdarlington4269]

it's interesting. I've never seen a video of yours before this, but I've definitely watched the other channels you've mentioned! Shows how networks work with algorithms. That said, it's not that surprising that a decent portion of your subscribers are subscribed to 1000+ channels- it's essentially the friendship paradox, right? You're more likely to be subscribed to by someone who subscribes to many people!

[@maxwibert]

@Not David is genuinely concerned that "ASMR" is dominating his node labels

[@helenfedorenko6419]

Thank you for such good content, quality, and entertainment. I discovered a Network theory and want to dive deeper into this topic. You have such a charismatic voice. It's so pleasurable to hear you. And your animation/visualization is the best.

[@not_David]

thank you for the kind words :)

[@shivvu4461]

9:15 I don't know how many channels I am subscribed to. But there are many channels which I used to watch few years ago and today I don't watch them. Still they're there in my subscription list. I never cared to unsubscribe them. So I think that's why it is common for many people to have thousand channels subscribed. But I think your content is worth a subscribe. (Thisbis first time your video came into my feed)

[@sum_rye_hash_321]

"Through the magic of rendering two of them" made me pause, write this, then subscribe lol

[@trevinbeattie4888]

I understood that reference!

[@KentoKei]

I love how, even though I haven't subscribed to (NOT) David, I did reconignize a lot of the channels (and are subscribed to a lot of them) that were shown in the large bubble distribution of your viewers I truly am a nerd... even if that means that I can bounce between history, gaming, physics, psychology, animation, sound design, and nature channels all within the time span of a single hour just because of how varied my subscriber list is from being on KZbin for so long and picking up, even if briefly, into a bunch of niches

[@ZephyrysBaum]

same lol, but add linguistics

[@great_icosahedron]

the technology connections bits absolutely destroyed me :'D

[@filipbambustal9957]

Thank you for putting the music in the describtion :D

[@ThePhoenixSlayer]

I don't mind being a fellow nerd, definetely not DAvid. It's nice being here with you.

[@Eulerstat]

This is fascinating to see. I’m a first year stat PhD student and network theory is likely what my dissertation will be in. I really appreciate this video since sometimes it’s easy to get distracted with dissimilarity/similarity matrices and forget the bigger picture of what network theory is doing.

[@WangleLine]

Every time I watch your videos I want to leave a comment about how lovely your animations are, but then I end up forgetting to do that, so, here you go!

[@Daniboid]

You had me at network analysis. I was hooked with the analogy to the applications for developmental neuroscience

[@mohammadfayez5036]

i have had this account for almost a decade now and yes i am subbed to almost a thousand channels because of how my taste changed and the number of channels i sub to in each niche. it seems natural to me

[@hildcit]

very much so! used the same account for over a decade, and naturally the number of subscribed to channels throughout the years have just increased. i used to try and clean out so to speak old channels i don't watch any more years and years ago, but honestly imo it's not necessarily worth it.

[@StarNova9]

I've definitely not seen such a fun and interesting video in a long time. It's relatable, I learned something, the animations were amazing and you have a great sense of humour. You really made something unique here and I loved every second of it

[@richtigmann1]

This is actually awesome. Its such a creative idea, and I love how you get statistics, graph theory, data science,p and a bunch of other stuff to represent all this data. It's such an awesome idea to use publicly available API data too, I love it. I bet this might not be too far from how KZbin's algorithms figure out how to recommend videos Now I really want to try this out for myself as well, and see what connections and representations I could make

[@TheArmyKnifeNut]

I guess i shouldnt be surprised that the algorithm recomended this to me when I watch or subscribe many of the channels I saw highlighted at 0:17. 😅

[@BroudbrunMusicMerge]

Amazing video! One note of interest: Because these are only the subscribers with public subscriptions, there may be a bit of selection bias in terms of demographics -- since, if I'm not mistaken, KZbin initially had public subscriptions be the default for new accounts, while now they're private for new accounts.

[@not_David]

Thank you! But definitely yes. Even ignoring that theres also the 'friendship paradox' bias (great video about it by Memeable Data). Essentially, people who sub a lot are going to overrepresented, while people who sub little are going to be underrepresented, even though we have strong reason to believe that in actuality people who sub very little are the vast majority of channels out there.

[@void6400]

Indeed summoning salt does get me hyped. That aside you deserve great success with the level of professionalism displayed with every video you've produced.

[@not_David]

thank you :)

[@carlosmachina]

Great video, awesome analysis and I'm sorry for being subscribed (at the moment) to 1001 different channels, and possibly adding to the toll on your computer's performance.

[@DaGrEeNtEaCoNnOiSsEr]

Wow, this is a great video for introducing network theory! I'm kind of a naive biologist that stumbled into network analysis recently and it's really cool to see how you handle the data. Also love how you threw in the am-i-a-nerd existential thought 😂

[@not_David]

I always appriciate the kind words from people doing really cool things with networks. Thank you :)

[@mobashshirkareem976]

KZbin should make a dating app where you can meet people who are ‘in your network’. The only issue is, our network is probably very male dominated.

[@geeknerd763]

Nice idea, but I have a feeling that it won't work well even with a balanced ratio between the genders because even those who aren't nerds will register and convolute the probability.

[@FunctionallyLiteratePerson]

Hey, queer people exist so I'm sure it might work out for some people! Though it probably is mostly straight men

[@NullHand]

Well, this will probably just accelerate the extinction of the human race. By most online metrics, we would have to date 2.something Bots between each real human. As technology marches on, the Bots will be AI tweaked at a faster than human evolution rate to maximise ”user retention”. Eventually we will just prefer the Bots. There might be a slim hope of redemption if the currently most successfull Human brood parasites doing business as ”Kittens” decides to stage a pre-emptive raid into all the major internet storage server racks riding on hijacked Roombas.

[@arobbo28]

The network's content might be male dominated. That doesn't necessarily mean that *interest* in the network is male dominated.

[@JoopPoop-si5cg]

Whats the issue? ":D"

[@spcraftsman2656]

I love how you *discretely* teach us and make us interested in math by exploring problems and ideas with various mathematical tools.

[@helenbrown7440]

The visual representation of the network at 10:38 looks like the millennium falcon

[@not_David]

you're not wrong...

[@h34dlez5]

your visualizations are insane 🤌🏾😂 great video

[@silverspuppet]

I am honestly amazed I haven’t stumbled upon this channel before. This is both so informative and so visually appealing I am confident it is trying to get me to attempt a similar style and waste way to much time trying to imitate this instead of making any progress on all the other projects I also started.

[@not_David]

its all about sticking to it! if you watch my first video it was nowhere near this polished. Its all about improving over time!

[@99seaweed]

Hey, you stumbled now and I guess that’s what the algorithm is for.

[@memeabledata]

So interesting! And I'm blown away by these animations

[@not_David]

I feel the same about your videos whenever I watch them haha. Maybe we have like a non-overlapping set of blender knowledge

[@memeabledata]

@@not_David Haha looks like it! I have no idea how you do most of your animations

[@adamhazelwood2694]

when you were talking about API's it got me to actually understand how some of that works, I kind of wish actual programming tutorials were like this that go in depth in a easily understandable way.

[@michaels.3709]

This was an awesome visualization of a topic (network theory) that I'm very interested to learn more about. As a direct consequence of this, I have subscribed.