I Made a Graph From My Subscribers to Prove You’re Nerds
Вставка
- Опубліковано 1 чер 2024
- 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 UA-cam network
14:20 The UA-cam 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... - Розваги
How many channels do you subscribe to?
57 most of them are specific and not one spammer. I guess I’m letting UA-cam to choose the content for me.
@@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.
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?)
I'm part of the 1000+ channels subscribed to
@@not_David I’ve mechanical engineering back ground and using UA-cam 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 UA-cam 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.
If you had to use network theory to realize you too are a nerd, you definitely are a nerd
I'm still in denial, likely a statistical fluctuation.
@@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
@@not_DavidAs a nerd who is subscribed to all top 10 channels you are one of us.
@@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.
You could try what I did and pretend to not be a nerd. But it doesn't work lol @@not_David
As a dishwashing powder user, "through the magic of rendering two of them" killed me lmao
Are you using too much dishwashing powder? :p
IM SO GLAD THIS WAS THE TOP COMMENT, I JUST DIED FROM THAT AND FEEL SO VALIDATED NOW LMAO
Same!
Good to see other people not skipping a step
And why is your car using red brake light turn as turn signals? Stop it!
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 😳
To quote a certain German bread: Nerd party!
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.
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.
"Do you want to enable personalized ad?"
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
So this is your attempt to not be a bot?
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.
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.
@@tisdaliyeah
@@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
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.
I genuinely feel for you wow
Y'all seriously need a union
You have my sympathies
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.
this is the bad place!!
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 UA-cam accounts for a *long* time would result in a lot of subscriptions
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.
Great Scott!
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
Hi, my name is @chopczyk374 and I have over 600 subscriptions. I'am also a youtube addict.
By the magic of making 2 of them
Lol
I don't like it, the system knows to much. I'm gonna have to go dark
He speaks r/oddlyspecific...
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
Wait - I wrongly assumed that my taste in UA-cam 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
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.
@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
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.
you should try visiting a makerspace, i couldn’t avoid this demographic if i tried.
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.
"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.
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.
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.
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.
same xD, I've been in youtube for 16 years though, so that is just a couple subscriptions a month
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)
@@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 UA-cam 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...
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
"You are not representative of the wider UA-cam 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.
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.
@@theaxer3751 I've heard of Mr. Beast. He was in a Mark Rober video.
@@theaxer3751yeah, opening YT on a vrowser I haven't signed in on is a terrifying barrage of nonsence
@@likebot.Sure it wasn't Mr. Beat? Or Mr. Bean?
@@caam0000 Were either of them working on the Team Trees project? I think that's where I heard of him.
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 :)
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
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.
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.
you guessed it haha (unfortunently)
@@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.
@@not_David oh no...
@@mehnemjeff4793gaming is for fun. Getting put to sleep is crucial for us insomniacs
@@not_DavidASMR is from hell and I am glad you agree 😂
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 UA-cam communities in the last portion. I hope this channel gets a lot more attention in the future - you certainly deserve it!
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...
@@not_David It would be lovely if we could poll your subscribers and see which ones have which versions of colorblindness :D
@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
to even think of this while making your video surely proves you are a great person@@not_David
What do you use for this? @@not_David
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.
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 :)
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.
Interesting! I'll look into that! Again keep up the great work mate@@not_David
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.
"Less mainstream" lol
I was an alcoholic, eating balls of mozzarella on a mattress _on the floor,_ watching videos of people *dying!*
and that's not all I use two UA-cam 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 UA-cam 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"
Kinda like survivorship bias
@@liam8370 yeah that when you know your like a *NERD* nerd, you use incognito not for rule 34 ow2, but for *tech videos*
@@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.
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 💚
my immediate reaction to your comment was to chuckle. i will now be _not_ thinking about how sad that sounds.
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
my god you're right
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)
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.
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
if i had to guess why james, itd probably just be theodd1sout stans :>
@@WoolyCow yeah thats exactly why, "Animation" was also in that community not to mention his podcast also in that community
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.)
my favourite comment so far
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.
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)
@@suncat530sameeee
good luck! im at 550+ , so its too late for me!
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!
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!!
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
I'll give you a hint, it starts with A and ends with R
@@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...
This comment needs more interaction, I had to scroll way too long to find it
someone else found it was asmr
@@leave-a-comment-at-the-door finally
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.
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.
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.
@@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.
@@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
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.
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!
Awesome video! The blender graphics compliment the video very well too, they are very smooth. Well done!
It's so fun watching this video and being like 'oh hey I know that channel'
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
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 UA-cam aren't all so random after all
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.
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.
Exactly
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
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
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
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
Never in a million years would I have guessed that Drawfee would’ve been featured in this video
just trying to contribue to that 2million grind
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 😋
Or even shawn!
Hi fellow Nerds! ❤
hi
I am in this comment and I appreciate it.
Hey!
I guess I am here too :(
@@not_David good place to be!
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
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!
thank you for the kind words :) (and on the uhmms video as well so I'm not spamming you lol)
I love that somewhere on UA-cam there is a Barbeque Gaming community
I get such whiplash when I see some of my friends’ UA-cam 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 UA-cam algorithm knows we like to watch the most.
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.
"... *I* might be a nerd" he said, looking up from the math homework he asssigned himself.
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
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. ❤
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
Same here
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
Hell yeah
@@ItsAsparageeseit does 😢, on a more serious note I have around 40ish playlists with all my interests and just rotate around
Seeing the tiny Drawfee bubble sandwiched between the titanic Veritasium and Tom Scott ones was really funny to me, for some reason
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!
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 :)
16:16 Bird. Bird is the 4-letter word, it must be.
got it in one!
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”?
The real network was the friends we made along the way
HAHA
Yes
That's true
aww :3
Many years of watching youtube made me sub to 1000+ (closer to 2000 lol) channels. Recently, I ran a browser script to remove all of them. Decided to become real picky with my subscriptions. You earned a sub!
I counted I have 986 subscriptions, I thought I had more. I only have 11 on my other account though.
Just found this UA-cam page, but the accuracy is amazing 😂, officially subscribed
I'll be honest and say I used to collect UA-cam 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!
im currently subbed to close to 5k channels lol
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
Just under 8k subscriptions for me
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
really well said
YT Alg decided to shake this video out onto my main page. Network videos/projects like this always tickle a special region in my brain it seems. So great vid.
The thing that really [weirdly] caught my attention though was the note you added at 12:07 about your voice sounding different. FYI, if you recorded one section at a different time during the session, then your voice was probably tired. NileRed in a podcast (Safety Third, maybe??) mentioned that this was something he noticed in his videos and it annoyed him a lot during the editing process. I don't recall how he addressed the issue, but at least its something to keep in mind.
This is an incredible video. I’m don’t know how I was recommended this, but now I’m so glad I was. Subscribed!
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?”
very nice comment, thank you :)
15:13
I think James refers to TheOdd1sOut
right group but why would his name appear?
He is that popular, lol@@florianschneider3982
I’ve been watching UA-cam for over a decade. I’ve been subscribed to thousands of channels because of how many different interests I’ve had in my time watching UA-cam. If there is even a single video I like, I tend to subscribe to see more stuff like that… but I’m also still subscribed to UA-camrs from longer ago. It’s really interesting to me how people can’t be subscribed to close to 1000 channels.
9:12
I'm not a bot, yet I'm running up against the upper limit of the number of subscriptions that UA-cam allows...
Ya how
[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.
990
85
I'm at 860.
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.
14
Honestly top tier UA-camr 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!
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.
Highly agree, you're punching well above your subscriber weight. Keep this up and you'll be at 1m in no time
I wish I was still in school. Data science is huge right now, and insanely interesting
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 💗
Awesome video, as always!
Your analysis prompted me to check my own subscribed channels. Good news first: it's not a thousand... 976, so everything is fine. As I scrolled through them, I recognized the overwhelming majority of them and usually remembered a few scenes immediately. My theory on why there are so many is that younger people's interests change frequently (I started watching UA-cam at 14), and the algorithm does its best to keep up by suggesting possible new interests. This increases the fluctuation in the channels you watch even more. When I found a "new thing," I always clicked through a lot of videos and channels. I often treated the subscription like a "like+" and "rewarded" impressive creations or complex videos with subscriptions, sometimes after a few minutes. I remember subscribing to channels even when I didn't watch the video through, but acknowledged the effort that was put into it and felt bad. Nowadays, I subscribe to maybe two channels per month or so. It seems that the accumulation of ever-changing interests over years is what leads to such big subscription numbers. If I were to start over now, I think it would take a long time to reach even 100.
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!
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 UA-cam'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 UA-cam that I'd never experienced before and wouldn't normally find through a typical recommendation network.
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
@@not_DavidDo you have the data available somewhere? I wonder what channels are involved in inter community connections
This is the first video of yours that has even ever come across my feed. So glad I came across it.
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.
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 UA-cam for so long and picking up, even if briefly, into a bunch of niches
same lol, but add linguistics
@Not David is genuinely concerned that "ASMR" is dominating his node labels
the thing with having a lot of subscribred channels is that i mainly watch the recomanded tab and from times to times unsubscribes, but mainly keeps all the random ones i used to watch
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.
thank you for the kind words :)
"Through the magic of rendering two of them" made me pause, write this, then subscribe lol
I understood that reference!
I don't mind being a fellow nerd, definetely not DAvid. It's nice being here with you.
i'm subbed to more than 2k channels, but your vids are truely special and i really enjoy them
very kind words, thank you :)
I'm subscribed to well over two hundred different UA-cam channels, but only roughly half of those post videos with any frequency or regularity. Many of those channels are dormant, or I don't actively watch all of their videos--I just check in with them occasionally, depending on what I need. And I actively prune my subscriptions to only the ones I believe are still valuable to me! So I'm inclined to believe that 10% of your subscribers may very well be subscribed to over a thousand channels.
Super interesting - I'm usually not a data guy but you made this very digestible
my favourite type of compliment, thank you
Sounds like you didn't know you were a data guy. ;)
^ this
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
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.
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.
15:18 ah yes my favorite community James(?)
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.
I'm not one of those sick, disturbing over-1000-subscriptions -people but as someone who is nearing 700, for me it just kind of happened. I made my account in 2009 and I've watched youtube pretty much daily, usually multiple hours a day. Most of the channels I've subscribed to either post videos rarely or stopped posting them at all so I don't even get that many new videos in my sub box in a day. I rarely unsubscribe, and when that happens it's usually a channel that I subscribed to years ago posting something again, reminding me that I'm not interested in that anymore. Anyway, you're one of the newer ones but also one of my new favorites, keep making more nerdy stuff!
I wish there was like a "time you subbed" thing. It would be interesting to track how peoples subscription patterns/preferences change over time.
And also thank you :)
@@not_David I believe there is actually a way to do that, at least for one's own subscriptions (there are 3rd party tools that do it), though I am not sure if it is available as a part of the publicly available API for getting info about other people's channels.
@@PureAsbestos interesting, I'll have to take a look into that
I wished they increased the number of people you can be subscribed to
I always thought my 450 subs was way too much, it's reassuring to see there's people with way more than that lol
This is my first one of your videos I watched and from looking at the top 10 Ik I need to sub
This is so validating for my channel taste. The whiplash I experienced when the network was basically all channels I subscribed to... lmao. Awesome video, congrats on 50k!
Also, I am subbed to 184 channels. I'm not a bot!
UA-cam 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.
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.
Hey, queer people exist so I'm sure it might work out for some people! Though it probably is mostly straight men
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.
The network's content might be male dominated. That doesn't necessarily mean that *interest* in the network is male dominated.
Whats the issue? ":D"
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, UA-cam initially had public subscriptions be the default for new accounts, while now they're private for new accounts.
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.
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!
oh my. i am glad that this video made its way to me. this is incredibly well made and you got another subscriber now
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.
its all about sticking to it! if you watch my first video it was nowhere near this polished. Its all about improving over time!
Hey, you stumbled now and I guess that’s what the algorithm is for.
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
Bro what an introduction to your channel happy to say you got a new sub, I can’t possibly refuse to support someone willing to give me a digestible amount of nerdy information while maintaining a humorous approach 😊
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 UA-cam'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
Not David will have a million subscribers within 12 months and I live for it. These videos are great! I once heard CGP Grey talk in a podcast about a myth that "youtube is full" but he argued that creators that are good story tellers and have interesting things to say will always be succesful, Not David is an example of this. I once commented on creating a high school maths lesson using pokemon trading cards, still working on it but it's damn fun at the moment!
thank you for the kind words
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.
thank you :)
I'm not subscribed to you and this was the first video of yours i saw but i am subscribed basically all the channels you highlighted so the algorithm definitely got the connection you outlined here when it recommended me this video lol
I'm one of those people who are subbed to almost 1000 channels. All I can say is, I've been on here for a while. You find a LOT of really cool people making excellent content including you.
I love how you *discretely* teach us and make us interested in math by exploring problems and ideas with various mathematical tools.
I almost never comment on videos anymore, but this one left such a lasting impact on me even after I was done watching it yesterday that I just had to come back and leave a comment. I've heard about how graph theory is useful for learning about many different kinds of things but I've never really realized the scope of that until I saw how it was applied in this video. It was also really nice to see how connected the UA-cam "nerd" community really is! A lot of the large mathy and science-y channels were to be expected, but it was really cool to see some of the smaller channels I watch also be included in the list (I was really surprised to see Junferno in there).
I guess what I'm trying to say is that I really liked the idea behind this video and I appreciate all the effort that must've gone into making it. I hope you remember me when this channel inevitably blows up!
(And also, do you have a Discord? It would be cool to be able to stay connected in some way, but I did see your latest community post mentioning it so I understand if that's not something you're interested in.)
I'm a big junferno fan so I'm glad they showed up in the network!
And thank you for the very kind comment :) Unfortunently as you noticed no discord.. but I'm still scrolling through these comments lol
This is the first video I've seen from this channel, and it convinced me to subscribe. I'm also using it to decide who I should watch next lol