What are some ‘power of friendship’ shows (anime or non-anime) you can recommend me? I’m looking for things to watch to recover from making this video lol. EDIT: for those of you wondering, the power of friendship of the brain is 5.11. I am bummed I didn't think to calculate that in the video -_-
oo I know nothing about Fairy Tail, is it power of friendship-y? I didn't get that vibe from the pictures I saw but obviously thats a poor way to judge lol
@@not_David Yes. It is basically all about power of Friendship. And it has ships......... though only occasionally the kind seen in One Piece ;) I should warn you that it's long, but compared to OP it is not.
What I'm getting from this video is that friendship depends on some time spent in proximity to each other. Yet, we've never spent any time together. I see how it is.
15:54 I take a break from working on my thesis proposal to watch a goofy anime video, and all of a sudden the video is about my thesis topic :0 Great video! That’s what I love about network science - asking how “things are related to other things” can reveal so many interesting similarities.
It's actually incredible how you achieved that I watched for 16 minutes without understanding anything, just until I got my educational part, thank you!🎉
@@Fenhum I definitely didn't think so but apperently? That's okay, everyone who has watched it has been incredibly receptive to it so I am okay with that :) Besides, the YT algorithm works in wierd ways -- the Uhmms video sat at 3K views for a month before the algorithm got to it and then it took off in a week lol.
8:17 there are 6 seasons of Jojo bizarre adventures, but the first season has both part 1 and 2, and part 3 is divided on 2 different seasons, and from there, each part had just one season
oh yeah for sure (though I've heard others argue there are more). But what I think is interesting is that the network agrees that part 1 and 2 should be classified as part of the same community, where as the splitting of part 3 into two seasons is a bit arbitrary (likely due to production) and that's also mirrored in the network as they are not split apart there.
I made graphs for supports for a couple of Fire Emblem games, and they can be pretty cool to look at. You can sometimes see little 'factions' emerge in the various armies.
Oooh I've only played the FE on the gameboy advanced (I am old -_-) but I feel like FE would be a really good series to do that on. You should make a video on it if you havn't already, I'd love to see it!
Ok I feel like I want to elaborate, because just typing "production quality gud" is a bit cheap: Everything about this video feels perfect to me. The script is very well thought out, even with some of the jokes scrapped (tho I'd like to hear them if I'm being honest). The visuals are obviously amazing, the style is a very pleasant mix between clean/professional and a certain artistic quality that fits the character of the entire work very well and that I personally really like. The sound design, while often overlooked in educational videos, is absolutely on point in my opinion. The music is incorporated nicely while not being overwhelming. And most importantly, especially for hearing-impaired folks like me: I can understand every word you say. That's obviously at least partly due to you speaking very clear and with a pacing that's easily understood by even non-native speakers. But beyond that I feel like the way you edited your voice (with compression, eq and what have you) is perfect for this kind of educational content. I see (or rather hear) more and more content creators in recent years process their voice to such an extent that there's almost only bass left, which might fit the character of some radio shows but it makes for a very unpleasant hearing experience for people who have hearing loss in the higher frequencies (upward of 1-1.5kHz, which is pretty much everyone with hearing loss). Ok so now I've went on quite the rant, which wasn't really my intention but as this issue is very near and dear to my heart I wanted to express my appreciation of how you approach your production. Also there are actual not-auto-generated English subtitles, which I just now noticed because I didn't at all need them in order to be able to understand you :D That makes the entire accessibility thing even better. So yeah, kudos to you and I'm very excited to keep supporting you and your work!
Amazing comment, thank you :) I do take accessability very seriously and I'm still not perfect at it, but I do my best. I think non-auto generated subtitles are very important because I can add a bit of emphasis or something to carry the tone across even if someone can't hear a thing. There is room for improvement though (especially with all my spelling mistakes haha). Again though, thank you :)
@not_David It takes a bit of fortitude to get through the show, it's got Old Anime syndrome with the drawn out power up scenes and fights The TYBW is crazy 11/10
This is so cool. I genuinely think i would have done better in math class if I was taught the things I could do rather than rote memorization of formulas. I made sure I understood ratios and percentages because I knew where to use those things IRL. I've been talking with various people recently about education and social media and I just hope videos like this get out there and inspire kids to learn.
my favourite kind of comments because this was it exactly for me. I did super poorly in high school because of this but it wasn't until I got to uni where my profs showed me how super creative math/science can be. I try to take that with me when making these videos. Thank you :)
My simplistic thought about the small network factor would be the main character focus that results in a few constant noies which are always involved for most connections (read: episodes) - would be interesting so see how a random network behaves when introducing such largely consitent nodes and if it meaninfully changes type. Maybe the same with the brain regarding the left-right corpus callosum interconnect?
That's what I originally thought and wasn't even going to do the small worldness because I thought it would be trivial because of that. However, because of the math I did to make the friendshipscores between 0 and 1 (details in description), and then the cutoff threshold, its actually no longer true -- the main character is often not the most popular character in the show because of this. So theres something else that is driving the smallness of the network.
While anime is not my cup of tea, you consistently kept my attention with absolutely great visuals, interesting theory and good humor throughout. The sudden switch to the brain to show the applicability of network theory was a real treat. I love the videos, keep them coming!
Thank you :) Part of the reason this video took so long to make was because I wanted to make sure the script did not alienate people who havn't watched anime. I'm glad to see the effort worked for a lot of people.
Amazing as usal! My favourite maths youtuber and 80k crossover was the best christmas present i could have gotten, an applied mathematician is long overdue on their podcast, hope this is them turning their heads towards that...
This video is so good what the hell Networks have been a fascination of mine for a little while and featured extensively in some personal projects of mine, but I've only ever actually covered network adjacent subjects in my education, so this stuff just tickles my brain endlessly. I could listen to you talk about networks all day :)
thanks so much :) I never had it pop up in my studies so when I was first introduced to it I was slightly intrigued but then someone soon after introduced me to a blog where the authour was trying to identify the main character of star wars using network theory and I was blown away. That blog inspired a lot of this video in fact (at least thematically).
@not_David aaand that's a Google search and a bookmark for some light reading later! Is it the one by Evelina Gabasova? Data science content is such a great niche on TH-cam, and everyone once in a while a nerdy-ass video will absolutely blow up, and I know you've got the sauce. Keep it up! I'll be on your patreon :)
Yes the one by Evaline Gabasova, it was really good! I'd also recommend Networks an Introduction by Mark Newmann if you can find it. The first third of that book is a math-free look at some of the things networks can be used for. Its from the early 2000s but it inspired the youtube subcriber video I made. Then it goes into all the math of how to do all that stuff.
I was under the impression that for JoJo - Season 1 is part 1 and 2 - Season 2 and 3 are both part 3 - Season 4 is part 4 - Season 5 is part 5 - Season 6 is part 6
Not sure how much of an impact this has large-scale, but it's interesting to me that this will give false friendship values for characters who appeared in the same episode together, even if they never appear in the same scene together - the most obvious example of this being A-plot characters sharing episodes with B-plot characters.
That is entirely a possibility. That's more or less what the cutoff is trying to account for. If two sets of characters appear together because of A-plot and B-plot stuff, and it happens once then they likely won't pass the threshold. If it keeps happening then that is often because they are in fact related somehow, but its not necessarily true. I admit though, I can't really think of an example where two long running side-by-side stories are occuring, but the characters between them are not aware of each other or don't become aware of each other... but I'm sure it happens.
Well, third comment on this video. I've dabbled in graph theory in a bioinformatics class and it's really cool to see similar concepts being explored on this channel. It may be another area you could look into for more graph algorithms and approaches if you're interested! The context for me was identifying protein-protein interaction clusters by seeing how often they come up together which is basically finding groups of friends in this context, very cool! I dropped the class for now so I'm actually not that suited to give you futher reading but I know there is more out there in this field! Cheers!
actually protein interaction networks is something ive really been wanting to look into a bit more. What little I've seen in that area has been mindblowingly complex
Absolutely phenomenal video, super interesting and an engaging way to learn. I study mechanical engineering but your channel has inspired me to take a graph theory elective in my fourth year!
Fantastic video. Very very well made. The topic it cool, the animations are well made, the jokes are super, the references makes me nostalgic, the integers shown with two decimal places precision are precise 🤷♂😂. I loved it. I might have to look up some network algorithms, the video made me think about how to do these computations not to think of doing them efficiently. 👍
I briefly woke up to get some water at 3am and saw this uploaded. I don't have the time to watch it for now, but just know I liked the vid and commenting to try and help it get recommended more. Can't wait to watch this later! :D
I love how, not only is this an incredibly cool and interesting video, but it also contains a really clear love of the "edutainment" community of video creators on TH-cam In terms of my own random associations, this feels a lot like a concept BDG would investigate in Unraveled, though the execution and presentation is still very much your own. Fantastic work
Oh, this video was so delightful as always! I have to admit, some of the brain stuff is slightly beyond my reach, but the basis of it all was explained so clearly I am fairly sure I still got the gist (and man it really is incredibly interesting) And as usual, the editing and animation are absolutely top notch - I loved the comic book style effects you chose (and it's always fun to see a Persona 5 cut-in, hehe)
Great video! Loved the animation, your application of the rigorous scientific process and the super neat way you taught us many of these techniques and your interpretation of the results Thanks for taking the time and effort to make this and putting in all this love into it! Had two follow up questions - - 1. Which characters were on the different longest path in the different shows? - 2. What was the Power of Friendship of the brain?
Thank you :) 1. Great catch! Its not unique. Theres actually quite a llootttt of characters whose distance is 5. I suspect that is due to the 'hub' characters like luffy who connect a lot of other nodes. I opened my graph and I think one example would be 'Charlotte Basskarte' and 'Shandia'. I actually have no clue who those are, but thats because the largest distances are statistically going to be between the characters that show up the least. 2. I was actually kicking myself the night before I didnt think of it. I actually just calculated it and its really low at 0.5. I kind of suspect that I maybe did something wrong in a rush, or its possible that I need to rethink the break threshold. The brain network is different in that you can have negative links (brain activity in one region is anti correlated with other regions). Thats very common and could mess up how I take the threshold value because in the anime network it can only be positive.
I usually don't write youtube comments... But I just wanted to tell how unbelievably good this video is. The graphic design, the animations, the science... I have absolutely no interest in anime, but I still loved every second of this video - well done!
Thank you! That genuinely means a lot to me. Part of the reason this took as long as it did to make was to make the script understandable/enjoyable even if you havn't watched the anime. There are plenty of in-jokes for those that do, but my goal was to make it transparent even if you don't (I succeed in some places better than others admittedly).
@@not_David The humor definitely worked for me! I got a lot of the jokes, and for others you could just guess the punch line by the way you wrote/delivered them. Great stuff!
Yes! Although im still catching up with the series (and actually making network spoiled some things for me... lets just say I thought T.D.Law was going to be like a one off side character in sabaody and then he kept popping up in my graph metrics -_- I don't know how he is involved and what he does but now I know hes certainly not a side character)
Wouldnt it make more sense to run the (non modularity) calculations for the communities aswell and then using the average of all those calculations as x and the calculation of the whole network as y and the final friendship score = modularity • x + (1-modularity) • y Feels like takes into account that very strong friendships might be present in a community but communities themselves might be more isolated from eachother. A handful of close friends is more powerful than a thousand acquaintances
I was thinking about doing something analogous to this. Or phrase differently -- should the score for jojo be the score for the whole network or the individual communities averaged together? My reasoning was that I wanted the 'power of friendship' to reflect the *anime*. Like if someone really likes 'power of friendship shows' and you know that the last season of a show has the most power of friendship you've ever seen, is it still a good recommendation for that person who would have to sit through a lot of 'non-power of firnedship' just to get to the good stuff? I don't think there is a right or wrong answer here, but I opted to treat the shows as a whole instead of as a sum of their parts.
Please do a follow up where you dive even deeper into the fun stuff since you went to all the trouble to gather all those data about the episodes in every character! The twist was really fun but I legit was sad that you weren't going to talk about the most fun clusters or connections you found in each, or trying to differentiate friends and enemies via some sort of faction tagging Anyway, loved the video!
Thank you! I was actually thinking of doing some indepth dives on a seperate channel (since as you said I already have them). I had a section in the video where I did do that but I felt it alienated people who didn't know the shows. Some of those things are possible, but the friends vs enemies thing is difficult. By which i mean, finding the 'sign' of a friendship is a genuinely difficult problem in social network analysis. Often it is very easy to find that there is a relationship but finding if its a positive or negative relationship is really difficult. I reckon i could also talk about that if I made a video like that.
It wasn't intentional haha I am treally trying to keep these shorter but the topics have become more complex (that's also why they take so long to make). But still, thank you :)
My thoughts: 0- Congrats for all the work ! 1- I'm surprised you did not picked Fairy Tail which is infamous for its power of friendship. 2- Also I'm pretty sure Zoro and Ace should share a connection (see the end of Alabasta) but I get why (the threshold). 3- I also see a big wich is "enemies". Let's say Luffy and Don Flamigo, according to the graph they should be good friends.... quite the opposite. 4- You should read Bleach, not watch nor drink it. 5- For the brain, what were the data? The cross "luminosity", the co-frequency, something else ?
Thanks for your thoughts! 0: Thank you :) 1: I didn't know anything about Fairy Tail but I've seen some people mention that as well, it's certainly much higher on my list now 2: Exactly, they don't show up together enough to get a link to pass the threshold 3: Yeah for sure. "Friends" here should probably be more like "are they aquiainted". Maybe this would be a first step to weed out all the connections to know who to focus on, but to get negative relations you'd need to go into the show most likely. That being said, in these shows the enemies do often turn out to be friends later on lol 4: Don't worry I always read the warning labels whenever handling Bleach 5: It was Blood Oxygen Level Dependant signals from fMRI, essentially just how oxygenated the blood is which is taken as a proxy for neural activity
i know you put a lot of effort into this, so i wanted to point out a couple of visual effects i loved :D 1:00 the dot matrix to the right 1:09 literally everything about this, the subtle wave, the chromatic aberration, however you changed the strength of the dots 1:37 the hat 3:12 i hope your computer enjoyed rendering that one 3:46 lovely compositing 5:37 the change in the shadow angle 6:50 car 9:39 just really nice 11:39 erdos feeling left out :( 12:10 the clouds 14:16 juicy ty for the great vid eric
This was such a nice comment, it left a small tear in my eye haha. I was having trouble sleeping (someone in the street below yelled and woke me up 6___6) and I was listening to the final fantasy 9 song 'you are not alone' randomly and was reading this and it was such a mood holy moly. Thank you :')
Fantastic video! As an anime fan and a fan of your videos, this was a super fun watch. I do wonder a bit about how you're scoring friendship strength. Particularly in that there are characters who have strong friendships as a foundational aspect of their origin story, but that are separated from these past friends and so we don't see them in the show but they are referenced. Also, to echo what I read in some comments here, a lot of the friendship building comes in the filler episodes. But regardless, this was an entertaining watch! And I learned some stuff I didn't already know about networks :)
Regarding your first point -- for sure, this is definitely a (very) rough approximation (hence 3:33). It's entirely possible that in this particular example (i.e., fictional characters) one could get a more nuanced view. But that would also raise as many questions as to how one should quantify the friendship score and I suspect it will introduce some subjectivity. And that is actually a real problem that real world social network researchers face, its just that topic in and of itself would be like a 30 minute video on its own lol
"Luckily, today's sponsor isn’t trying to sell you anything." Yeah, they are selling something, but "when you do things [well], people won’t be sure you’ve done anything at all." Ironically, what they are selling is incompatible with the power of friendship. (To be clear, I deeply respect you. You have arete and compassion.)
so what is the not David Power of Friendship index of the brain network? For note 1: One idea I have is to divide N by max(M1, M2) or min(M1, M2), where Mi is the number of episodes between the first episode that character i appeared and the latest episode of the show. However, that will result in an unequal treatment of one-off characters that appear later on in the show vs one-off characters that appear early on in the show, so it's probably not that great of an idea.
this is actually what I did originally! There are ways to augment it but inevitably I ran into the same problem -- when you have main characters (like naruto) that show up in the vast majority of the episodes, they just get connected to everyone, and especially with the intervals they end up getting really strong connections to people they met like 2 or 3 times.
10:42 I'm unsure if something was lost in translation with links, but didn't zoro meet Ace? As in they were in an episode together and also physically met? Surely that means the link would be one and not two?
Great question! It's because of the threshold. They meet, but they don't meet often enough to pass the threshold, where as Luffy and Ace do meet enough to have a link between them. Admittedly they don't physically meet in a lot of them (e.g., marineford, impel down) but appearing in the same episodes tends to suggest they are somehow related which is what we're seeing.
Ah of course I understand now. Would I then be right in saying that in the MRI example this would be represented by some neurons firing together but not frequently enough to be considered connected?
@ exactly! The brain has a whole lot of activity and sometimes two neurons will just happen to fire together even though they have nothing to do with each other. This is to get rid of those types of accidental links
I had a philosophy class teacher who had published a paper on exactly that, about some tricks on how to make them and an analysis of the results on french science-fiction. It's in french, though: Yannnick Rochat and Mathieu Triclot, 2017. In any case, glad to see more interesting in this technique :D
time to put my canadian french to practice! Thank you for sharing I'm always interested in similar work to this because I can see how I could have done it differently/better
Independant from your analysis: You should watch Bleach 🙂 Bleach has a relative big cast of main protagonists that spent a lot of time together (epsiodes), the bad boys build always groups and hang around, and there are not many one-time characters. I guess that is what shows in the graphs.
I do really want to. I watched it on YTV (canadian network) a long time ago, but when it got to about half way through the soul society arc towards the begining of the series YTV reset from the begining (they did this a lot I think to try to give time for more material to come out, they did it to inuyasha like 5 times). I do genuinely think Bleach has the coolest style of the 5 shows (aside from maybe jojo, or at least tied). The original OP was asthetically a huge inspiration for me in making this video (though then Dan Da Dan came out and the style pivoted more towards that lol).
Interesting you choose the compliment of modularity as a friendship score. I would have been inclined to use the value directly. The close-knit communities in jo-Jo would seem to imply they are closer friends than the homogeneous bleach “anyone goes”
Yeah that's a great point and I debated this a lot as well. I really can't recall what it was that made me choose the compliment. There was a reason for it (likely because of the Jojo example I gave) but I admit the alternative is also very persuasive.
I NEED you to make a character map for the show Arcane. From a writer's perspective, that show is incredibly well-written and executed due to it's strong characterization and worldbuilding - largely thanks to established lore from the video game League of Legends. I predict that it would have a high number of paths and interconnectivity, without too much clustering.
I think arcane would be interesting to look at pre and post time skip (s1, I haven't yet watched s2). It would be really cool to see the changes there I think. However, I think for arcane you'd want to maybe take a more hands-on approach because it is so much smaller. The entire point of the shows I chose was that its effectively not feasible to do it manually so we need alternative approaches. But that doesn't mean that those approaches should always be used and I think Arcane is an example where I'd want a more in-depth look.
I'm glad you enjoyed it :) I put a lot of effort into trying to write the script to not alienate people who haven't seen the shows so I'm super happy to hear that.
Algorithm really screwing you over rn This whole topic seems really overwhelming, how would you even get into this field of study? Great video, and as expected, great explanations as well!
Thank you :) I highly recommend 'networks an introduction' by mark newman. Its a classic and starts off with just some examples of things network theory could be used for. Typically making the network is the hard part though and its often very data specific. That just comes from working in the field.
12:10 At least in shonen anime, it's pretty easy to imagine how the small world network occurs: The story usually focuses on a specific protagonist character, whom we see the world through from start to finish, so the longest paths will often converge in this main character. Simply put, the protagonist has two friends, so the parents of those friends are only 4 steps separated from each other even if they never actually met in the story. Maybe something like Hunter X Hunter, who has functionally multiple protagonists, can produce a larger network.
Exactly, I think that's part of the reason as well. Even in Jojo you do have one or two characters, especially Dio and Jotaro and Joseph who span a couple of seasons which is enough to cause it as well. I think HxH is a great idea to test. I have seen it but it didn't occur to me to test it.
7:27 I think for Dragon Ball it’s because there’s only like 3 characters that play a role in every part of Dragon Ball. Quick count is Goku, obviously, Krillin, and Bulma. There are arguments to be made for characters like Roshi, Yamcha, and Chi Chi, who appear in nearly every part of the series, but they end up not being important later on so their appearances are more treated like cameos. While shows like One Piece and Naruto create a cast that survive through each arc and do important things in each arc keeps their numbers higher.
hmm interesting. I noticed your example used only Dragon Ball characters but I do include DBZ, DBGT, DBS (daima came out after I did my analysis). Were you just using those as an example or do you think it extends generally?
@@not_DavidI was just using Dragon Ball characters because they are the only ones that appear in Dragon Ball, Z, and Super. People like Vegeta or Trunks don’t appear in the original Dragon Ball, it’s a backwards compatibility thing. Goku, Krillin, and Bulma go through the entire series more or less each acquiring the same connections, so they’re kinda treated like one character. If people like Yamcha and Tien kept up in fighting strength with Goku, I think it would be more like One Piece or Naruto, where they have a big team that are important the entire time. It’s like if Naruto’s characters like Rock Lee and Might Guy fall off in power and would’ve not been able to make any impact in the Ten Tails arc.
@@friggy1899 ahhh sorry yes I see now. Yeah I think that's probably not too far off the mark. For me my reasoning was, I think, somewhat similar in spirit. While Jojo is very explicitly seperated into parts, I think dragon ball does it too, and I think what you're saying is tied to that -- a lot characters are introduced but often not carried forward, which from a network point of view makes it very modular.
Hi, it's me again, I still don't watch Anime. I found this video relatively easy to understand, but I'm still kind of a nerd, so take my opinion with a grain of salt
thank you :) I was a bit bummed at first but I've been really enjoying the supportive messages people have been leaving. It's been its own kind of power of friendship thing lol
On accounting for the truangular structure: you could divide by the numver of episodes after both characters are introduced. This would change it from a measure of how important the friendship is to the character in the most episodes to how important the friendship is to the show after it's possible to show it. E.g. if Luffy makes a super friend 75% of the way through the show and they never split up it makes sense for it to be 1 not .25, and this removes the "single episode of perfect friendship" issue
hmm I have to think about it some more but on initial glance I like it. I did try something similar by looking at only the windows where the character with less episodes appears. This is kind of like what you suggest except I stop looking after they leave the series which doesnt solve the problem because the main character is just so likely to appear in the episode. But I think by continuing to count to the end of the series like your method would do kinda solves that problem.
@not_David if your main goal is to filter out "main character syndrome" then you could instead divide by the priduct of their episode counts, that way minor characters would have strobg friendships witg each other, but very weak friendships with main characters, this wssentially flips what I was suggesting before. But if you choose which thing you are trying to measure, you can use that to define which metric is best suited.
Yeah that's also a really good suggestion. Someone suggested doing N^2/(N1N2) where N = shared episodes, N1 and N2 are number of episodes for each character. I like this because its very similar to other measures people use (or example cross correlation). I will say though, that the current method does actually acomplish in getting rid of main character syndrome in the same was as you describe it, but I the product method because it is more standard and analytically tractable (no max() function). I think the product rule does solve it better though because the minor character has influence on the score where the way I do it, only the major character influences the score.
At one point in the history of the script I did do a more indepth dive into each network one by one but I felt it really would only make sense to viewers who knew the anime, where as I was trying to write this so it would be understandable even without having seen the shows. I was thinking of maybe going a bit more indepth into them on the second channel but we'll see if I have the time
What are some ‘power of friendship’ shows (anime or non-anime) you can recommend me? I’m looking for things to watch to recover from making this video lol. EDIT: for those of you wondering, the power of friendship of the brain is 5.11. I am bummed I didn't think to calculate that in the video -_-
Fairy Tail
oo I know nothing about Fairy Tail, is it power of friendship-y? I didn't get that vibe from the pictures I saw but obviously thats a poor way to judge lol
@@not_David Yes. It is basically all about power of Friendship. And it has ships......... though only occasionally the kind seen in One Piece ;)
I should warn you that it's long, but compared to OP it is not.
Also as you might have heard My Little Pony: Friendship is Magic is also about it. Though people for some reason hate the fandom.
I mean you can watch like the opposite of power of friendship which is like jjk but I’m not really sure
What I'm getting from this video is that friendship depends on some time spent in proximity to each other. Yet, we've never spent any time together.
I see how it is.
hey hey hey not true, case and point 3:35
are you two friends?
well the english youtube maths sector aint that big
i need to know how the Show My Little Pony: Friendship is magic holds up compared to these xD
15:54 I take a break from working on my thesis proposal to watch a goofy anime video, and all of a sudden the video is about my thesis topic :0
Great video! That’s what I love about network science - asking how “things are related to other things” can reveal so many interesting similarities.
Ooo what are you doing for your thesis?
Graph models of the brain!
very cool, both of you! 👍
Yes he’s back!
6 months for a 20 minute video -_- what a ratio haha
@@not_Davidwith quality like this, we're happy to wait!
Welcome back , "David. "
If that's even who you are, "David"
Lol
It's actually incredible how you achieved that I watched for 16 minutes without understanding anything, just until I got my educational part, thank you!🎉
This video is criminally underrated. I wonder why it has little views?
*Wink* *Wink* *checks the connection between you and anime youtubers...*
haha thank you :')
@@not_David Is my reasoning accurate to the real data?😅
If so, we anime-science nerds might be a rare specimen.
@@Fenhum I definitely didn't think so but apperently? That's okay, everyone who has watched it has been incredibly receptive to it so I am okay with that :) Besides, the YT algorithm works in wierd ways -- the Uhmms video sat at 3K views for a month before the algorithm got to it and then it took off in a week lol.
Another anime-science nerd reporting for duty! o7
@not_David I see. Good luck then!
@@BooLightning Oh so you're an anime-science nerd? Name every sci-fi anime.
Friendship is Witchcraft was not a reference i was expecting but by God I am here for it lmao
8:17 there are 6 seasons of Jojo bizarre adventures, but the first season has both part 1 and 2, and part 3 is divided on 2 different seasons, and from there, each part had just one season
oh yeah for sure (though I've heard others argue there are more). But what I think is interesting is that the network agrees that part 1 and 2 should be classified as part of the same community, where as the splitting of part 3 into two seasons is a bit arbitrary (likely due to production) and that's also mirrored in the network as they are not split apart there.
I made graphs for supports for a couple of Fire Emblem games, and they can be pretty cool to look at. You can sometimes see little 'factions' emerge in the various armies.
Oooh I've only played the FE on the gameboy advanced (I am old -_-) but I feel like FE would be a really good series to do that on. You should make a video on it if you havn't already, I'd love to see it!
@not_David oh boy... videos are easy to watch but hard to make hahaha
today is indeed a great day to see a breakdown breakdown
Your production quality is crazy, this video was so worth the wait! And also worth disobeying you and paying more for your Patreon :P
Ok I feel like I want to elaborate, because just typing "production quality gud" is a bit cheap:
Everything about this video feels perfect to me. The script is very well thought out, even with some of the jokes scrapped (tho I'd like to hear them if I'm being honest).
The visuals are obviously amazing, the style is a very pleasant mix between clean/professional and a certain artistic quality that fits the character of the entire work very well and that I personally really like.
The sound design, while often overlooked in educational videos, is absolutely on point in my opinion. The music is incorporated nicely while not being overwhelming. And most importantly, especially for hearing-impaired folks like me: I can understand every word you say. That's obviously at least partly due to you speaking very clear and with a pacing that's easily understood by even non-native speakers. But beyond that I feel like the way you edited your voice (with compression, eq and what have you) is perfect for this kind of educational content. I see (or rather hear) more and more content creators in recent years process their voice to such an extent that there's almost only bass left, which might fit the character of some radio shows but it makes for a very unpleasant hearing experience for people who have hearing loss in the higher frequencies (upward of 1-1.5kHz, which is pretty much everyone with hearing loss).
Ok so now I've went on quite the rant, which wasn't really my intention but as this issue is very near and dear to my heart I wanted to express my appreciation of how you approach your production. Also there are actual not-auto-generated English subtitles, which I just now noticed because I didn't at all need them in order to be able to understand you :D That makes the entire accessibility thing even better.
So yeah, kudos to you and I'm very excited to keep supporting you and your work!
Amazing comment, thank you :) I do take accessability very seriously and I'm still not perfect at it, but I do my best. I think non-auto generated subtitles are very important because I can add a bit of emphasis or something to carry the tone across even if someone can't hear a thing. There is room for improvement though (especially with all my spelling mistakes haha). Again though, thank you :)
You got me over here mumbling to myself about the joys of Bleach and then the video switches to cool MRI stuff, not expected but very cool
I will watch bleach! ... one day... I hear really good things about TYBW (and the animation looks amazing)
@not_David It takes a bit of fortitude to get through the show, it's got Old Anime syndrome with the drawn out power up scenes and fights
The TYBW is crazy 11/10
great video, now lets watch it!!! I love your videos man!!
Thank you, i hope it was worth the wait!
This is so cool. I genuinely think i would have done better in math class if I was taught the things I could do rather than rote memorization of formulas. I made sure I understood ratios and percentages because I knew where to use those things IRL. I've been talking with various people recently about education and social media and I just hope videos like this get out there and inspire kids to learn.
my favourite kind of comments because this was it exactly for me. I did super poorly in high school because of this but it wasn't until I got to uni where my profs showed me how super creative math/science can be. I try to take that with me when making these videos. Thank you :)
The new visualization look so good
this reminds me of that one anon that solved a permutation problem due to an anime argument
I vaguely remember hearing about this... I need to look it up again...
The network guy is back!
My simplistic thought about the small network factor would be the main character focus that results in a few constant noies which are always involved for most connections (read: episodes) - would be interesting so see how a random network behaves when introducing such largely consitent nodes and if it meaninfully changes type. Maybe the same with the brain regarding the left-right corpus callosum interconnect?
That's what I originally thought and wasn't even going to do the small worldness because I thought it would be trivial because of that. However, because of the math I did to make the friendshipscores between 0 and 1 (details in description), and then the cutoff threshold, its actually no longer true -- the main character is often not the most popular character in the show because of this. So theres something else that is driving the smallness of the network.
Thank you very much for the detailed description box! Highly appreciate the work that is put into it (and espacially the sources incl. Music).
Thank you! I spend maybe too long writing those sometimes and its nice seeing people going into them and reading it haha
While anime is not my cup of tea, you consistently kept my attention with absolutely great visuals, interesting theory and good humor throughout. The sudden switch to the brain to show the applicability of network theory was a real treat. I love the videos, keep them coming!
Thank you :) Part of the reason this video took so long to make was because I wanted to make sure the script did not alienate people who havn't watched anime. I'm glad to see the effort worked for a lot of people.
6:57 Talk about eye candy.
that should have been the thumbnail lol
biggest plot twist of the video tbh
Super cool video, as I could way easier follow with the face validity I can muster.
Excellent for showcasing networking! You're doing an awesome job!
Amazing as usal! My favourite maths youtuber and 80k crossover was the best christmas present i could have gotten, an applied mathematician is long overdue on their podcast, hope this is them turning their heads towards that...
Thank you haha :) I appriciate the kind words
This video is so good what the hell
Networks have been a fascination of mine for a little while and featured extensively in some personal projects of mine, but I've only ever actually covered network adjacent subjects in my education, so this stuff just tickles my brain endlessly. I could listen to you talk about networks all day :)
thanks so much :) I never had it pop up in my studies so when I was first introduced to it I was slightly intrigued but then someone soon after introduced me to a blog where the authour was trying to identify the main character of star wars using network theory and I was blown away. That blog inspired a lot of this video in fact (at least thematically).
@not_David aaand that's a Google search and a bookmark for some light reading later! Is it the one by Evelina Gabasova?
Data science content is such a great niche on TH-cam, and everyone once in a while a nerdy-ass video will absolutely blow up, and I know you've got the sauce. Keep it up! I'll be on your patreon :)
Yes the one by Evaline Gabasova, it was really good! I'd also recommend Networks an Introduction by Mark Newmann if you can find it. The first third of that book is a math-free look at some of the things networks can be used for. Its from the early 2000s but it inspired the youtube subcriber video I made. Then it goes into all the math of how to do all that stuff.
I love how the video is really educational but so well disgused with the anime theme, amazing work!
I was under the impression that for JoJo
- Season 1 is part 1 and 2
- Season 2 and 3 are both part 3
- Season 4 is part 4
- Season 5 is part 5
- Season 6 is part 6
You tricked me into learning about neuroscience using anime! Wow what a great video.
Not sure how much of an impact this has large-scale, but it's interesting to me that this will give false friendship values for characters who appeared in the same episode together, even if they never appear in the same scene together - the most obvious example of this being A-plot characters sharing episodes with B-plot characters.
That is entirely a possibility. That's more or less what the cutoff is trying to account for. If two sets of characters appear together because of A-plot and B-plot stuff, and it happens once then they likely won't pass the threshold. If it keeps happening then that is often because they are in fact related somehow, but its not necessarily true. I admit though, I can't really think of an example where two long running side-by-side stories are occuring, but the characters between them are not aware of each other or don't become aware of each other... but I'm sure it happens.
Well, third comment on this video. I've dabbled in graph theory in a bioinformatics class and it's really cool to see similar concepts being explored on this channel. It may be another area you could look into for more graph algorithms and approaches if you're interested! The context for me was identifying protein-protein interaction clusters by seeing how often they come up together which is basically finding groups of friends in this context, very cool! I dropped the class for now so I'm actually not that suited to give you futher reading but I know there is more out there in this field! Cheers!
actually protein interaction networks is something ive really been wanting to look into a bit more. What little I've seen in that area has been mindblowingly complex
Wow this video is a masterpiece 🔥
It's amazing that videos like this are available for free!
that's high praise wow, thank you
BRUH NO SHOT YOU HAVE IMPOSTER SYNDROME. You have some of the highest quality vids. You are underrated. I'm shellshocked
Thank you for the very kind words lol :)
Absolutely phenomenal video, super interesting and an engaging way to learn. I study mechanical engineering but your channel has inspired me to take a graph theory elective in my fourth year!
That's such a nice comment, thank you I am honoured :)
What amazing production quality with interesting topics!
Thank you, I'm glad you enjoyed it :)
Fantastic video. Very very well made. The topic it cool, the animations are well made, the jokes are super, the references makes me nostalgic, the integers shown with two decimal places precision are precise 🤷♂😂. I loved it. I might have to look up some network algorithms, the video made me think about how to do these computations not to think of doing them efficiently. 👍
The visuals are insane bro nice!
Another Not David video just as I get back home from work. Life is good.
I briefly woke up to get some water at 3am and saw this uploaded. I don't have the time to watch it for now, but just know I liked the vid and commenting to try and help it get recommended more.
Can't wait to watch this later! :D
haha thank you! This video is not worth healthy sleep, it will be there for you later!
I love how, not only is this an incredibly cool and interesting video, but it also contains a really clear love of the "edutainment" community of video creators on TH-cam
In terms of my own random associations, this feels a lot like a concept BDG would investigate in Unraveled, though the execution and presentation is still very much your own. Fantastic work
any mention of my work in the same sentence as BDG is an immense compliment thank you :)
@@not_David It was certainly meant as one!
Oh, this video was so delightful as always! I have to admit, some of the brain stuff is slightly beyond my reach, but the basis of it all was explained so clearly I am fairly sure I still got the gist (and man it really is incredibly interesting)
And as usual, the editing and animation are absolutely top notch - I loved the comic book style effects you chose (and it's always fun to see a Persona 5 cut-in, hehe)
Those are very kind words, thank you :)
I love this channel bro
You really outdid yourself with this video.
yay graph theory man :D keep going dude
I didnt even know I was subscribed to you!
Thank you for uploading captions at the same time :) I truly appreciate it
Watching the video, I'm sure it'll be great, keep up the good work 👍
Thank you :) I hope you enjoy it
@not_David great video 😄 thank you for your hard work. as always, the visuals are amazing
Great video! Loved the animation, your application of the rigorous scientific process and the super neat way you taught us many of these techniques and your interpretation of the results
Thanks for taking the time and effort to make this and putting in all this love into it!
Had two follow up questions -
- 1. Which characters were on the different longest path in the different shows?
- 2. What was the Power of Friendship of the brain?
Thank you :)
1. Great catch! Its not unique. Theres actually quite a llootttt of characters whose distance is 5. I suspect that is due to the 'hub' characters like luffy who connect a lot of other nodes. I opened my graph and I think one example would be 'Charlotte Basskarte' and 'Shandia'. I actually have no clue who those are, but thats because the largest distances are statistically going to be between the characters that show up the least.
2. I was actually kicking myself the night before I didnt think of it. I actually just calculated it and its really low at 0.5. I kind of suspect that I maybe did something wrong in a rush, or its possible that I need to rethink the break threshold. The brain network is different in that you can have negative links (brain activity in one region is anti correlated with other regions). Thats very common and could mess up how I take the threshold value because in the anime network it can only be positive.
oh no, I just calculated the formula wrong. It is actually 5.11 so very high power of friendship!
Huzzah!
It takes the whole gang to make it through each day, so this number definitely tracks
Thanks for your answers!
I usually don't write youtube comments... But I just wanted to tell how unbelievably good this video is. The graphic design, the animations, the science...
I have absolutely no interest in anime, but I still loved every second of this video - well done!
Thank you! That genuinely means a lot to me. Part of the reason this took as long as it did to make was to make the script understandable/enjoyable even if you havn't watched the anime. There are plenty of in-jokes for those that do, but my goal was to make it transparent even if you don't (I succeed in some places better than others admittedly).
@@not_David The humor definitely worked for me! I got a lot of the jokes, and for others you could just guess the punch line by the way you wrote/delivered them. Great stuff!
I had no idea you were a one piece fan! I absolutely adore the show and the little references are amazing!
Yes! Although im still catching up with the series (and actually making network spoiled some things for me... lets just say I thought T.D.Law was going to be like a one off side character in sabaody and then he kept popping up in my graph metrics -_- I don't know how he is involved and what he does but now I know hes certainly not a side character)
@ I had marineford spoiled for me I was extremely mad lol
Worth waiting 6 months for lol Great Video!
Wouldnt it make more sense to run the (non modularity) calculations for the communities aswell and then using the average of all those calculations as x and the calculation of the whole network as y and the final friendship score = modularity • x + (1-modularity) • y
Feels like takes into account that very strong friendships might be present in a community but communities themselves might be more isolated from eachother. A handful of close friends is more powerful than a thousand acquaintances
I was thinking about doing something analogous to this. Or phrase differently -- should the score for jojo be the score for the whole network or the individual communities averaged together? My reasoning was that I wanted the 'power of friendship' to reflect the *anime*. Like if someone really likes 'power of friendship shows' and you know that the last season of a show has the most power of friendship you've ever seen, is it still a good recommendation for that person who would have to sit through a lot of 'non-power of firnedship' just to get to the good stuff? I don't think there is a right or wrong answer here, but I opted to treat the shows as a whole instead of as a sum of their parts.
Please do a follow up where you dive even deeper into the fun stuff since you went to all the trouble to gather all those data about the episodes in every character!
The twist was really fun but I legit was sad that you weren't going to talk about the most fun clusters or connections you found in each, or trying to differentiate friends and enemies via some sort of faction tagging
Anyway, loved the video!
Thank you! I was actually thinking of doing some indepth dives on a seperate channel (since as you said I already have them). I had a section in the video where I did do that but I felt it alienated people who didn't know the shows.
Some of those things are possible, but the friends vs enemies thing is difficult. By which i mean, finding the 'sign' of a friendship is a genuinely difficult problem in social network analysis. Often it is very easy to find that there is a relationship but finding if its a positive or negative relationship is really difficult. I reckon i could also talk about that if I made a video like that.
Loved the section about brains!
Man, I saw a trend of increasing video length from you lol. Good work on this video too! Really enjoyed the content
It wasn't intentional haha I am treally trying to keep these shorter but the topics have become more complex (that's also why they take so long to make). But still, thank you :)
19:25 guy really went "quantum entanglement, quantum Schrodinger theory, quantum temporal anti-negentropy."
nice video, apple juice is better than orange juice the people have spoken
excuse me but if you check the poll, orange juice is like 3 pixels ahead of apple juice.
My thoughts:
0- Congrats for all the work !
1- I'm surprised you did not picked Fairy Tail which is infamous for its power of friendship.
2- Also I'm pretty sure Zoro and Ace should share a connection (see the end of Alabasta) but I get why (the threshold).
3- I also see a big wich is "enemies". Let's say Luffy and Don Flamigo, according to the graph they should be good friends.... quite the opposite.
4- You should read Bleach, not watch nor drink it.
5- For the brain, what were the data? The cross "luminosity", the co-frequency, something else ?
Thanks for your thoughts!
0: Thank you :)
1: I didn't know anything about Fairy Tail but I've seen some people mention that as well, it's certainly much higher on my list now
2: Exactly, they don't show up together enough to get a link to pass the threshold
3: Yeah for sure. "Friends" here should probably be more like "are they aquiainted". Maybe this would be a first step to weed out all the connections to know who to focus on, but to get negative relations you'd need to go into the show most likely. That being said, in these shows the enemies do often turn out to be friends later on lol
4: Don't worry I always read the warning labels whenever handling Bleach
5: It was Blood Oxygen Level Dependant signals from fMRI, essentially just how oxygenated the blood is which is taken as a proxy for neural activity
3:04 that low-key gave me flashbacks to the percolation video 😂
haha I was thinking the same as I was making it
The visuals are so technically impressive ❤!!!
Thank you :)
yay! not david is back
the video quality is insanely good brother, the content as usual top notch. loved every second
Thank you, I'm glad you enjoyed it :)
The breakdown breakdown in the beginning got me
i know you put a lot of effort into this, so i wanted to point out a couple of visual effects i loved :D
1:00 the dot matrix to the right
1:09 literally everything about this, the subtle wave, the chromatic aberration, however you changed the strength of the dots
1:37 the hat
3:12 i hope your computer enjoyed rendering that one
3:46 lovely compositing
5:37 the change in the shadow angle
6:50 car
9:39 just really nice
11:39 erdos feeling left out :(
12:10 the clouds
14:16 juicy
ty for the great vid eric
This was such a nice comment, it left a small tear in my eye haha. I was having trouble sleeping (someone in the street below yelled and woke me up 6___6) and I was listening to the final fantasy 9 song 'you are not alone' randomly and was reading this and it was such a mood holy moly. Thank you :')
Fantastic video! As an anime fan and a fan of your videos, this was a super fun watch.
I do wonder a bit about how you're scoring friendship strength. Particularly in that there are characters who have strong friendships as a foundational aspect of their origin story, but that are separated from these past friends and so we don't see them in the show but they are referenced.
Also, to echo what I read in some comments here, a lot of the friendship building comes in the filler episodes.
But regardless, this was an entertaining watch! And I learned some stuff I didn't already know about networks :)
Regarding your first point -- for sure, this is definitely a (very) rough approximation (hence 3:33). It's entirely possible that in this particular example (i.e., fictional characters) one could get a more nuanced view. But that would also raise as many questions as to how one should quantify the friendship score and I suspect it will introduce some subjectivity.
And that is actually a real problem that real world social network researchers face, its just that topic in and of itself would be like a 30 minute video on its own lol
@@not_David that would be a 30min video I'd watch! Lol 😆
"Luckily, today's sponsor isn’t trying to sell you anything." Yeah, they are selling something, but "when you do things [well], people won’t be sure you’ve done anything at all." Ironically, what they are selling is incompatible with the power of friendship.
(To be clear, I deeply respect you. You have arete and compassion.)
how dare this only have 7631 views :C
Bro literally pulled the inside out twist
This is an amazing video, thank you so much.
Thank you :) I'm glad you enjoyed it
The fact this video has less than a million views is simply outrageous.
This video is so good ! ❤️
Thank you! I'm glad you enjoyed it :)
so what is the not David Power of Friendship index of the brain network?
For note 1: One idea I have is to divide N by max(M1, M2) or min(M1, M2), where Mi is the number of episodes between the first episode that character i appeared and the latest episode of the show. However, that will result in an unequal treatment of one-off characters that appear later on in the show vs one-off characters that appear early on in the show, so it's probably not that great of an idea.
this is actually what I did originally! There are ways to augment it but inevitably I ran into the same problem -- when you have main characters (like naruto) that show up in the vast majority of the episodes, they just get connected to everyone, and especially with the intervals they end up getting really strong connections to people they met like 2 or 3 times.
finally
my algorithm knows what I want to watch
Negative memory and rest is so relatable
it took me a while to figure out what this was in reference to and now I get it and I agree haha
10:42 I'm unsure if something was lost in translation with links, but didn't zoro meet Ace? As in they were in an episode together and also physically met? Surely that means the link would be one and not two?
Great question! It's because of the threshold. They meet, but they don't meet often enough to pass the threshold, where as Luffy and Ace do meet enough to have a link between them. Admittedly they don't physically meet in a lot of them (e.g., marineford, impel down) but appearing in the same episodes tends to suggest they are somehow related which is what we're seeing.
Ah of course I understand now. Would I then be right in saying that in the MRI example this would be represented by some neurons firing together but not frequently enough to be considered connected?
@ exactly! The brain has a whole lot of activity and sometimes two neurons will just happen to fire together even though they have nothing to do with each other. This is to get rid of those types of accidental links
Commenting for engagement before watching the video. And well ... for friendship
the power of friendship :')
Man
What a cool video
Damn it
Insane quality
I had a philosophy class teacher who had published a paper on exactly that, about some tricks on how to make them and an analysis of the results on french science-fiction. It's in french, though: Yannnick Rochat and Mathieu Triclot, 2017.
In any case, glad to see more interesting in this technique :D
time to put my canadian french to practice! Thank you for sharing I'm always interested in similar work to this because I can see how I could have done it differently/better
My god, what an excellent video.
Thank you :) I really appriciate it
really nice video
Thank you :)
Every time you post it's a masterpiece! Impressive!
That is very kind of you, thank you :)
Independant from your analysis: You should watch Bleach 🙂
Bleach has a relative big cast of main protagonists that spent a lot of time together (epsiodes), the bad boys build always groups and hang around, and there are not many one-time characters. I guess that is what shows in the graphs.
I do really want to. I watched it on YTV (canadian network) a long time ago, but when it got to about half way through the soul society arc towards the begining of the series YTV reset from the begining (they did this a lot I think to try to give time for more material to come out, they did it to inuyasha like 5 times). I do genuinely think Bleach has the coolest style of the 5 shows (aside from maybe jojo, or at least tied). The original OP was asthetically a huge inspiration for me in making this video (though then Dan Da Dan came out and the style pivoted more towards that lol).
Interesting you choose the compliment of modularity as a friendship score. I would have been inclined to use the value directly. The close-knit communities in jo-Jo would seem to imply they are closer friends than the homogeneous bleach “anyone goes”
Yeah that's a great point and I debated this a lot as well. I really can't recall what it was that made me choose the compliment. There was a reason for it (likely because of the Jojo example I gave) but I admit the alternative is also very persuasive.
What I learned from this video is that I fried my brain by watching anime.
I NEED you to make a character map for the show Arcane. From a writer's perspective, that show is incredibly well-written and executed due to it's strong characterization and worldbuilding - largely thanks to established lore from the video game League of Legends. I predict that it would have a high number of paths and interconnectivity, without too much clustering.
I think arcane would be interesting to look at pre and post time skip (s1, I haven't yet watched s2). It would be really cool to see the changes there I think.
However, I think for arcane you'd want to maybe take a more hands-on approach because it is so much smaller. The entire point of the shows I chose was that its effectively not feasible to do it manually so we need alternative approaches. But that doesn't mean that those approaches should always be used and I think Arcane is an example where I'd want a more in-depth look.
As some one who hasn't seen any of these shows this is so interesting
I'm glad you enjoyed it :) I put a lot of effort into trying to write the script to not alienate people who haven't seen the shows so I'm super happy to hear that.
Algorithm really screwing you over rn
This whole topic seems really overwhelming, how would you even get into this field of study?
Great video, and as expected, great explanations as well!
Thank you :)
I highly recommend 'networks an introduction' by mark newman. Its a classic and starts off with just some examples of things network theory could be used for. Typically making the network is the hard part though and its often very data specific. That just comes from working in the field.
ok this was veeery unexpected
I hope in a good way lol
12:10 At least in shonen anime, it's pretty easy to imagine how the small world network occurs: The story usually focuses on a specific protagonist character, whom we see the world through from start to finish, so the longest paths will often converge in this main character. Simply put, the protagonist has two friends, so the parents of those friends are only 4 steps separated from each other even if they never actually met in the story. Maybe something like Hunter X Hunter, who has functionally multiple protagonists, can produce a larger network.
Exactly, I think that's part of the reason as well. Even in Jojo you do have one or two characters, especially Dio and Jotaro and Joseph who span a couple of seasons which is enough to cause it as well. I think HxH is a great idea to test. I have seen it but it didn't occur to me to test it.
7:27 I think for Dragon Ball it’s because there’s only like 3 characters that play a role in every part of Dragon Ball. Quick count is Goku, obviously, Krillin, and Bulma. There are arguments to be made for characters like Roshi, Yamcha, and Chi Chi, who appear in nearly every part of the series, but they end up not being important later on so their appearances are more treated like cameos. While shows like One Piece and Naruto create a cast that survive through each arc and do important things in each arc keeps their numbers higher.
hmm interesting. I noticed your example used only Dragon Ball characters but I do include DBZ, DBGT, DBS (daima came out after I did my analysis). Were you just using those as an example or do you think it extends generally?
@@not_DavidI was just using Dragon Ball characters because they are the only ones that appear in Dragon Ball, Z, and Super. People like Vegeta or Trunks don’t appear in the original Dragon Ball, it’s a backwards compatibility thing. Goku, Krillin, and Bulma go through the entire series more or less each acquiring the same connections, so they’re kinda treated like one character. If people like Yamcha and Tien kept up in fighting strength with Goku, I think it would be more like One Piece or Naruto, where they have a big team that are important the entire time. It’s like if Naruto’s characters like Rock Lee and Might Guy fall off in power and would’ve not been able to make any impact in the Ten Tails arc.
@@friggy1899 ahhh sorry yes I see now. Yeah I think that's probably not too far off the mark. For me my reasoning was, I think, somewhat similar in spirit. While Jojo is very explicitly seperated into parts, I think dragon ball does it too, and I think what you're saying is tied to that -- a lot characters are introduced but often not carried forward, which from a network point of view makes it very modular.
I was here befor the algorithm.
The full metal alchemist and fusion joke was cursed
Hi, it's me again, I still don't watch Anime. I found this video relatively easy to understand, but I'm still kind of a nerd, so take my opinion with a grain of salt
Thank you for letting me know :)
let's all love linear algebra
sorry that the algo screwed you over. great video.
thank you :) I was a bit bummed at first but I've been really enjoying the supportive messages people have been leaving. It's been its own kind of power of friendship thing lol
HAHA! Jokes on you I'm studying neuroscience at university and love this video!
duped at my own game :(
On accounting for the truangular structure: you could divide by the numver of episodes after both characters are introduced.
This would change it from a measure of how important the friendship is to the character in the most episodes to how important the friendship is to the show after it's possible to show it.
E.g. if Luffy makes a super friend 75% of the way through the show and they never split up it makes sense for it to be 1 not .25, and this removes the "single episode of perfect friendship" issue
hmm I have to think about it some more but on initial glance I like it. I did try something similar by looking at only the windows where the character with less episodes appears. This is kind of like what you suggest except I stop looking after they leave the series which doesnt solve the problem because the main character is just so likely to appear in the episode. But I think by continuing to count to the end of the series like your method would do kinda solves that problem.
@not_David if your main goal is to filter out "main character syndrome" then you could instead divide by the priduct of their episode counts, that way minor characters would have strobg friendships witg each other, but very weak friendships with main characters, this wssentially flips what I was suggesting before. But if you choose which thing you are trying to measure, you can use that to define which metric is best suited.
Yeah that's also a really good suggestion. Someone suggested doing N^2/(N1N2) where N = shared episodes, N1 and N2 are number of episodes for each character. I like this because its very similar to other measures people use (or example cross correlation). I will say though, that the current method does actually acomplish in getting rid of main character syndrome in the same was as you describe it, but I the product method because it is more standard and analytically tractable (no max() function). I think the product rule does solve it better though because the minor character has influence on the score where the way I do it, only the major character influences the score.
Hey algorithm, this is engagement.
I'm kinda sad I didn't get to explore those network a bit more.
Like,.I'm curious of the results.
At one point in the history of the script I did do a more indepth dive into each network one by one but I felt it really would only make sense to viewers who knew the anime, where as I was trying to write this so it would be understandable even without having seen the shows. I was thinking of maybe going a bit more indepth into them on the second channel but we'll see if I have the time
W David post
Algorithm, please push this high-effort, high-value video.
Well I don't know about high-value hahah but thank you :')