...the product being sold.

November 9, 2011 6:54 AM

The social graph is neither. Maciej Cegłowski, owner/founder/operator/sole employee of Pinboard (recently), blogger of idle words, lays down some science and thoughts about the charting of your personal connections, and why it's doomed.
posted by ardgedee (78 comments total)

I *just* finished reading this in another tab — brilliant from start to finish.
posted by mathowie at 6:58 AM on November 9, 2011 [2 favorites]

Thrown a sheep?
posted by Wolfdog at 7:03 AM on November 9, 2011 [1 favorite]

I've never been a fan of XFN. It's nice that such an awesome guy writes such a great article that I agree with so much.
posted by Plutor at 7:06 AM on November 9, 2011 [1 favorite]

I love how he gets into how stupid and arbitrary XFN is. Which is what MeFi uses for its contacts system, of course!
posted by zsazsa at 7:07 AM on November 9, 2011 [2 favorites]

Read it last night. The whole thing is fantastic.
posted by immlass at 7:14 AM on November 9, 2011

Yeah, the first time I heard about XFN I just laughed. A classic case of misapplied ontology.
posted by DU at 7:14 AM on November 9, 2011

"No problem," says Poindexter. "We'll add a time series of state transitions and exponentially decaying edge weights, model group dynamics as directional flows, and pass a context object in with each query..." and around we go.

I remember being in RDF discussions where this would always come up as a hand-wavey solution whenever the real world reared its ugly head.

This obsession with modeling has led us into a social version of the Uncanny Valley

This is a great way of putting it. Our current models give 2% very interesting recommendations and the rest are split between super obvious and completely off the mark.
posted by stp123 at 7:27 AM on November 9, 2011

To miss the point a bit: of course it's a graph; it just has labels on the edges. That's a trivial generalization.
posted by Jpfed at 7:28 AM on November 9, 2011 [1 favorite]

Asking computer nerds to design social software is a little bit like hiring a Mormon bartender.

If library schools had looked for their academic respectability in classification rather than third-rate computer science, they would be well-positioned to be relevant....

Now tell me one bit of original culture that's ever come out of Facebook.

To be fair, the endless pestering about Farmville is a kind of cultural experience.
posted by GenjiandProust at 7:28 AM on November 9, 2011 [4 favorites]

Yes, the name 'friends' was always just marking dribble, twitter got it right by naming the relationship 'following', maybe listening to, maybe stalking, maybe family but not listening or friends, etc., all represented by just simply 'following'.
posted by jeffburdges at 7:31 AM on November 9, 2011 [2 favorites]

Jpfed: "To miss the point a bit: of course it's a graph; it just has labels on the edges. That's a trivial generalization."

Except then you want to start adding metadata to the label, and then you get something that loosely resembles an edge, but also resembles a node with two edges. Faced with this problem, the folks behind UML invented the Association Class, which I've always found a somewhat amusing capitulation.
posted by vanar sena at 7:37 AM on November 9, 2011 [1 favorite]

I am constantly bemoaning the shift on facebook from Making New Friends to Connecting With Everyone You Know Already. in fact, I haven't really found a place on the internet that I was able to make friends on - interaction was either too sparse (metafilter), the userbase was too big (reddit), or we didn't have enough in common.

Old facebook was cool, back when you could search through your school and find people who had similar interests, or who were in the same classes. That's how I met my now fiancee, and several other good friends. Now, the only people Facebook will let me "connect" with are my grandfather's cousin, my childhood babysitter, people who I went to church with 10 years ago who I can't stand, some people out in Cali who are friends of friends but whom I will never ever meet (I'm not even really "friends" with our mutual friends, not since college anyway), and people who I already know but, let's face it, there's a good reason I'm not friends with them on facebook.

So this article really spoke to me. I really agree with the bit at the end about how what comes after this will make this look like garbage. I really hope so.
posted by rebent at 7:41 AM on November 9, 2011 [7 favorites]

Nailed it. Between this and his essay on interacting with fandom, Maciej is turning into my favorite guy on the internet whose name I can't pronounce.
posted by pts at 7:48 AM on November 9, 2011 [2 favorites]

Very nice post and article. I'm struggling with the direction to take my business' social media efforts in 2012 and this article just reinforces a lot of my gut feeling about what to do.

Best LOL line of the linked article:

"Asking computer nerds to design social software is a little bit like hiring a Mormon bartender. "
posted by webhund at 7:51 AM on November 9, 2011 [1 favorite]

If you're not following him on Twitter, you're doing it wrong. Terrific post, thanks. Also:

"One big sticking point is privacy. Do I really want to find out that my pastor and I share the same dominatrix?"

"While other dating sites nagged you to upgrade to an expensive 'Gold' status, which branded you as a foreveralone pariah . . ."
posted by yerfatma at 7:54 AM on November 9, 2011

I signed up with Pinboard for the excellent bookmarking service, but feeling proud for supporting someone who seems to be emerging as an excellent and influential blogger is icing on the cake. If you like this sort of writing, brows his archives... he's written several fascinating posts.
posted by gilrain at 7:54 AM on November 9, 2011 [1 favorite]

Yes, this was nice. I like cogent explanations like this that combine real thinking about the social implications of technology along with relatively detailed explanations of the problems with the technology at the level of implementation.
posted by OmieWise at 7:55 AM on November 9, 2011 [1 favorite]

Excellent article, and it conveys in a couple of sentences my feelings toward this bullshit called "social networking":

"Social networks exist to sell you crap. The icky feeling you get when your friend starts to talk to you about Amway, or when you spot someone passing out business cards at a birthday party, is the entire driving force behind a site like Facebook."

We have a name for the kind of person who collects a detailed, permanent dossier on everyone they interact with, with the intent of using it to manipulate others for personal advantage - we call that person a sociopath.”

posted by sutt at 7:56 AM on November 9, 2011 [9 favorites]

After reading this, as well as his pieces on Roman Polanski, French school lunches and "Dating without Kundera", I must publicly confess to a crush on Maciej Cegłowski.
posted by peacheater at 8:00 AM on November 9, 2011 [1 favorite]

I think this FPP couldn't be more wrong about its conclusion even though the linked blog post is right about every point. The thing is that there is no need for these tools to accurately model the complexity of our social reality. All they need in order to succeed is to be is useful.
posted by srboisvert at 8:15 AM on November 9, 2011 [4 favorites]

Yup, total mancrush here as well (even though he didn't answer a support email I sent to Pinboard once) and I'm annoyed that I didn't realise he'd transferred his main blogging efforts to blog.pinboard.in from Idlewords.

Other classics on Idlewords are Attacked by Thugs and Dabblers and Blowhards.
posted by Busy Old Fool at 8:18 AM on November 9, 2011 [1 favorite]

Along with thousands of others, I signed up for Pinboard as soon as it was made clear that Del.icio.us was EOL, and I have to say, I like being a customer of a website (this can't be a surprising position to take here on MeFi). And I know that Pinboard got another onslaught of signups right after the youtube guys relaunched Delicious. And I know that some of those people came to Pinboard with the desire to reproduce their Delicious friend-network there.

Because Maciej seems to be responsive to his customer base, I'm sure that he's been thinking about what/how/whether to implement social-network features, and I have to imagine this post was the product of that thinking. There's a lot of truth in it.

But as much as I have misgivings about Facebook, I have to admit that they're onto something. There is some value in being able to easily get in touch with everyone you know, to keep track of people you kinda-sorta know who you met at that thing that one time, etc. Even if the way we represent those relationships in a computer is unsatisfactory. I really want Diaspora—or something—to work and render Facebook and G+ irrelevant.
posted by adamrice at 8:18 AM on November 9, 2011

I am constantly bemoaning the shift on facebook from Making New Friends to Connecting With Everyone You Know Already.

Which is funny, because I feel like how FB uses the social graph is to say "all your buddies are friends with this person, so don't you want to be too?". The downside of this is that the people who are most likely to be friends with most of your friends (using the word in the FB sense and not in the real world sense) are people you do already know and choose not to "friend" on Facebook. So the people you get told to add all the time are the ones you really don't want to add.

Twitter definitely has this right. So does Dreamwidth, which uses a livejournal source code but separates access circles from reading circles. Even FB has caught a clue with public subscriptions and the "acquaintance" (low access "friend") model. The fact that these services model the complexity of our relationships a little better is part of what makes them more useful than, say, G+.
posted by immlass at 8:22 AM on November 9, 2011 [1 favorite]

The hard part about Facebook is managing it. They process incredible amounts of data, and the only way to pay for that is to turn the users into the product, which is a big part of this excellent article's main complaint. So whatever comes next needs to offload that part. Maybe the next 'social network' is a protocol, not a website.
posted by cell divide at 8:26 AM on November 9, 2011

"The hard part about Facebook is managing it. They process incredible amounts of data, and the only way to pay for that is to turn the users into the product, which is a big part of this excellent article's main complaint."

Wait, wasn't "customer as product" the whole point from the very beginning? The rest was just emergent properties of trying to optimize the moneymaking process.
posted by sutt at 8:31 AM on November 9, 2011 [1 favorite]

This is the best metafilter post I have seen in at least a week.

The money shot from the blog for me:

The funny thing is, no one's really hiding the secret of how to make awesome online communities. Give people something cool to do and a way to talk to each other, moderate a little bit, and your job is done. Games like Eve Online or WoW have developed entire economies on top of what's basically a message board. MetaFilter, Reddit, LiveJournal and SA all started with a couple of buttons and a textfield and have produced some fascinating subcultures. And maybe the purest (!) example is 4chan, a Lord of the Flies community that invents all the stuff you end up sharing elsewhere: image macros, copypasta, rage comics, the lolrus. The data model for 4chan is three fields long - image, timestamp, text.

He captures a lot of rant that I personally feel. Social graphs are tools. In the hands of trained and experienced sociologists they have been useful in the discovery of a number of new sociology facts, ideas, theories, &c. In the hands of people who are not trained and experienced they are almost invariably a waste of time and resources. It produces stuff like that idiotic OKCupid blog where they think their graph is telling them something about the state of sex relations in America, or the the state of race relations in major American cities, or what people really value in spite of what they say they value.

I don't mean to pick on that idiotic OKCupid blog; it is only one member of a population of thousands.

Do you ever look on page 4 or 5 or 6 of what Amazon.com's correlation algorithm thinks would interest you? I have, and it ain't even on the same planet let alone in the same ballpark. They probably don't see any problem at all with giving Mister Marcelus's wife a foot massage.

I had never heard of this guy before and will absolutely be looking at at least a couple more of those linked posts.
posted by bukvich at 8:56 AM on November 9, 2011 [1 favorite]

Maciej is very bright. This analysis is very good, but it's also a bit weird to me that anyone would ever think a graph database could actually capture social relationships. Of course it's more complicated than that! But it needs to be said, and Maciej said it well.

I want to drop a link here to Friend of a Friend Finder, a social graph research project I worked on back in 1997. Actually I did almost nothing; the idea was Pattie Maes' and I hated the project, didn't want to work on it. (Turns out she was right about it being an important idea.) I kind of still hate it, actually, the idea of quantifying social relationships in software just feels gross to me. Even if it is useful.
posted by Nelson at 9:01 AM on November 9, 2011 [1 favorite]

If the social graph is crude oil, doesn't that make our friends and colleagues the little animals that get crushed and buried underground?

Nah, no bias there.

The funny thing is, no one's really hiding the secret of how to make awesome online communities. Give people something cool to do and a way to talk to each other, moderate a little bit, and your job is done.

What about technological progress? I found the whole essay incredibly naive. Too many words to say "you are being sold." Yes, we know that. That's why we are careful with the information we share. Sure, the "social graph" is of dubious value ... to marketers. We know that.

However, users are finding it incredibly valuable and don't care too much about it marketers get their profiles wrong.

Now tell me one bit of original culture that's ever come out of Facebook

"Old facebook was cool, back when you could search through your school and find people who had similar interests, or who were in the same classes. That's how I met my now fiancee"

emphasis mine. i suppose you can define "culture" in many ways. i would count marriage as a cultural product.
posted by mrgrimm at 9:03 AM on November 9, 2011 [3 favorites]

Of course it's more complicated than that! But it needs to be said

Did it really? No digg.
posted by mrgrimm at 9:03 AM on November 9, 2011

In the old country, for example, we have two kinds of 'friendship' (distinguished by whether you address one another with the informal pronoun) and going from one status to the other is a pretty big deal; you have to drink a toast with your arms all in a pretzel and it's considered a huge faux pas to suggest it before both people feel ready

Can someone confirm or deny this? it is kinda cool.
posted by Ad hominem at 9:04 AM on November 9, 2011

I'm not sure I recognize the position he's arguing against as something that's being espoused much anywhere. What I've assumed people generally think of social graphs as generally implement is 'good enough for some purposes'.

But maybe I'm just missing some hype.
posted by Anything at 9:10 AM on November 9, 2011 [2 favorites]

and this one is absolutely fucking priceless:

"I am not qualified to call bullshit on Paul Graham when he writes about programming, history, starting a business, or even growing up as a social pariah, but I do know enough about art to see when someone is just making shit up. "
posted by bukvich at 9:13 AM on November 9, 2011 [2 favorites]

implemented
posted by Anything at 9:14 AM on November 9, 2011

This is the best article on social networks that I've read in a long, long time. It's so damn refreshing and such a relief that somebody out there with influence gets it.

I think all the recent changes with google have had me in a seemingly endless funk of perplexity.
posted by iamkimiam at 9:21 AM on November 9, 2011 [1 favorite]

It produces stuff like that idiotic OKCupid blog where they think their graph is telling them something about the state of sex relations in America

I don't know about that. Their sample size blows away pretty much any other sociology study ever.
posted by the jam at 9:23 AM on November 9, 2011 [1 favorite]

Ad hominem: In the old country, for example, we have two kinds of 'friendship' (distinguished by whether you address one another with the informal pronoun) and going from one status to the other is a pretty big deal; you have to drink a toast with your arms all in a pretzel and it's considered a huge faux pas to suggest it before both people feel ready

Can someone confirm or deny this? it is kinda cool.

I'm curious, too. I believe his old country is Poland, if that helps.
posted by gilrain at 9:25 AM on November 9, 2011

Can someone confirm or deny this? it is kinda cool.

I'm curious, too. I believe his old country is Poland, if that helps.

It's true. When Poles switch from using the formal form of address (pan/pani, literally sir/madam, as in 'Czy pan ma czas?' 'Does Sir have the time?') to the informal ty, they sometimes toast this transformative moment in their friendship by drinking a toast with their arms intertwined.

"As the relationship between the interlocutors develops over time, they may decide to cross the in-group/out-group boundary by changing address forms from pan(i) to ty. Traditionally, such a switch requires the Brudershaft ceremony, which has nearly died out among younger generations of Poles"
Lucja Biel, Emotional and social distance in English and Polish unpublished Ph.D. dissertation, 2004, Gdańsk

There are other ways that this shift is handled in other languages and they have been discussed before on a well-known Q&A site.
posted by Busy Old Fool at 9:27 AM on November 9, 2011 [10 favorites]

Well, while Maciej Cegłowski is right, there is one thing he is missing. The "social graph" isn't just about selling information to marketers - it's about selling the "social graph" to marketers. The consumer - in this case, advertisers - are being duped as well. And they aren't going to want to hear how their fancy schmancy tools are inadequate, for the same reason that the idiots on Wall Street don't want to hear it either.

For once, we sardines will have the last laugh.
posted by Xoebe at 9:28 AM on November 9, 2011 [3 favorites]

Globes are useless. The resolution is terrible, they don't zoom at all, you can't see current weather conditions, tides or even something as simple as local time of day.

1) The map is not the territory.

2) It's an iterative process not a destination.

3) We've only just begun iterating.
posted by scalefree at 9:30 AM on November 9, 2011 [8 favorites]

Normally reading something that expends so much effort tearing an idea down comes across as depressing, but I found it weirdly inspiring.

There is obviously a space for something new here that isn't facebook or g+.
posted by empath at 9:31 AM on November 9, 2011

Wolfdog: "Thrown a sheep?"

Just click the damn thing and leave my personal life out of it, OK pal?
posted by Reverend John at 9:31 AM on November 9, 2011

It produces stuff like that idiotic OKCupid blog where they think their graph is telling them something about the state of sex relations in America

It does tell you something about their user base. How well their user base matches up with "America" is another story. Either way, they're always interesting.
posted by empath at 9:32 AM on November 9, 2011 [1 favorite]

But one thing you can do is mine a huge amount of information about my friends and infer things about their interests, income, social status and tastes. And then maybe you can use that information to bring them valuable news and offers, or help them digitally engage with their favorite brands.

Exactly. It's not a graph, it's a spam database (much more boring), with the 'community' and users themselves as the data entry drones.

The Web (and Web conferences) are full of these 'misapplied ontologies' (DU, above).
posted by carter at 9:41 AM on November 9, 2011

Is G+' circles thing a way to get me to define and label my own relationships?
posted by ctmf at 9:48 AM on November 9, 2011

Google knows when I've added ardgedee to a circle called "best buds" and it knows when he's added me to a circle called "assholes" and it has much more about our relationship, and the way we think about each other, than we'll ever know ourselves.

There are really huge flaws in all of these social networks that make me just not give a shit about any of them except maybe Twitter (which has a workaround) but I don't care to describe them. I'm tired of having ideas on how to make things better and no time/money/headspace to see them through. Where are my minions? I have no minions. FLY MY MONKEYS
posted by seanmpuckett at 9:56 AM on November 9, 2011

Following on from my comment above, a further and for me more interesting twist in Polish is that there are two main words for the English 'friend': 'przyjaciel' and 'kolega'. (some argue there are three, but I don't think 'znajomy' really translates as 'friend', although I know it's not quite 'acquaintance' either) Note that this doesn't map onto the formal/informal forms of address I mentioned.

'Kolega' is perhaps the closest to the English 'friend', but you wouldn't expect much of a 'kolega' in terms of loyalty or support and it's almost impossible to lose that status by behaviour; it's much more about the fact that you know someone. 'Przyjaciel' is a term used for the 1/2/3/4 closest people in your life. Perhaps it has some equivalence with 'bestie'? Poles tend to feel that English-speakers label people as 'friends' too easily and to some extent offer good will too freely. A Pole is more likely to believe that there are very few true friends in life.

As an illustration, the frequency over one million words of the terms above (including feminine equivalents in Polish):
friend 262
kolega 242
przyjaciel 132
znajomy 46
colleague 19
acquaintance 7
Caveat: I couldn't find any information on the quality of the corpora used to compile this.

Note how there's basically one main word in English, with a few rarely-used alternatives, whereas with Polish the different words are used more often, probably leading to the need to make choices about displaying friendship status more often. The fact that Polish (also Russian and probably other similar languages) speakers have different linguistic and mental categories for these concepts is probably one of the reasons why Maciej can see that attempts to codify human relationships in XFN are likely to be very unsatisfying. And that's just Slavs - don't get me started on Korean concepts of friendship...

Sorry, this is probably stretching the boundaries of relevance to the article, but I find it terribly interesting.
posted by Busy Old Fool at 10:15 AM on November 9, 2011 [5 favorites]

Eh, I think the guy is being obtuse or is falling into the common programmer line of thinking that if the entire process can't be automated then the entire process is invalid.

You just need to give edges arbitrary weights. 1=twitter follower, 10=family member (all we need to know is the user picked spouse, the computer doesn't need to know what a spouse is, the human interperating the results knows what a spouse is) everyone else lies inbetween. Everone who picks lulzy relationships is chaff.

Coumputer can spit out reports all day long. Average weight of relationships, Aggregate weight, how many connections. It will just be up to a human analyst to make some deeper sense of it if that is what you want.
posted by Ad hominem at 10:18 AM on November 9, 2011

> that idiotic OKCupid blog where they think their graph is telling them something about the state of sex relations in America, or the the state of race relations in major American cities, or what people really value in spite of what they say they value.

OKCupid's scraping questionnaire responses and publishing analyses of them isn't really the same thing as trying to construct a social network model. For that matter, I enjoy reading when it's addressing stereotypes and how their observations contradict them (or reinforce them, for that matter). It's sociology lite and I don't think they make any claims of it being more than that.

Most marketing research, on which considerably large, expensive corporate decisions are made, is conducted with far more money and far less rigor.
posted by ardgedee at 10:30 AM on November 9, 2011

Woot! Recognition at last! I programmed a big part of the sheep throwing thing! Wrong side of history, here i come!

My grandchildren will think of me the way someone now thinks about yheur grandfather having been in the American nazi party.
posted by Ayn Rand and God at 10:33 AM on November 9, 2011 [3 favorites]

> To miss the point a bit: of course it's a graph; it just has labels on the edges.

He had bought a large map representing the sea,
Without the least vestige of land:
And the crew were much pleased when they found it to be
A map they could all understand.

"What's the good of Mercator's North Poles and Equators,
Tropics, Zones, and Meridian Lines?"
So the Bellman would cry: and the crew would reply
"They are merely conventional signs!

posted by jfuller at 10:35 AM on November 9, 2011

Well, I liked most of what the essay said, but it suffered from a common problem with tech writing which the need for people to proclaim "everything you know about topic X is wrong" its really annoying and one reason why I don't like to read tech blogs all that much.

Let's start with the basics. The problems he goes into are problems with the social graph not being an ontology not a graph. All you need for a graph are nodes (dots) and edges (lines between them) that's it. That's a mathematical graph.

I think, for people who have been using messageboards for years and years, facebook (at least in its current form, with your mom and all your friends from highschool) offers nothing. But for a lot of people it's their first experience with getting information and news from their friends. So for them, it has a lot of utility, because there not really interested in talking about random topics with random people. (Or ar least people who seem random from the start)

I think facebook runs the risk of becomming lame if the try to 'extract' value like crude oil. The metaphor is pretty creepy and sums up a lot of what people don't like about corporate social networing. But at the same time, while its unnerving for people like us, it's still true. But its more like fish stocks. If you try to extract all the value, you'll just kill it.but if you're smart you can take some profit maximizing amount -- pissing off the gnu hippies but leaving all gap-shopping consumerbots and their farmville farms yearning to digitally engage withtheir favorite brands.
posted by delmoi at 10:39 AM on November 9, 2011 [1 favorite]

It will just be up to a human analyst to make some deeper sense of it if that is what you want.

I think part of what he's saying is that the underlying data is essentially 'senseless' for many purposes beyond establishing basic demographic buckets for marketing purposes. This lack of sense in the underlying data arises in turn because social media attempts to impose a few rule-based categories on what is very subtle and complex human behaviour (what you think of the people in your life).

It's probably a mistake to think that the categories chosen by users are necessarily accurately indicative of what they feel about those people in their lives (for example).
posted by carter at 10:41 AM on November 9, 2011 [1 favorite]

The problems he goes into are problems with the social graph not being an ontology not a graph. All you need for a graph are nodes (dots) and edges (lines between them) that's it.

Well I think he was complaining that social networks aren't truely representable by a graph.
posted by empath at 10:42 AM on November 9, 2011

Does anybody know of a succinct critique of what google and facebook and amazon do with graph theory written by somebody with the expertise of, say, Wasserman and Faust?

This guy does a great job echoing my prejudices back towards me but I would be interested in seeing a similar piece written by somebody who could write the textbook, not just another guy who has read the textbook.
posted by bukvich at 10:51 AM on November 9, 2011 [1 favorite]

scalefree is correct.

Of course they are both social and graphs. The fact they're imperfect or don't do what you want them to do has nothing to do with what they are. Simulations don't need to be perfect models in order to be useful.

He may have legitimate points about what FB, Twitter, G+, etc. should be, but his bullet points are just silly.
posted by CheeseDigestsAll at 10:55 AM on November 9, 2011 [4 favorites]

Let me just start off by saying that I'm one of those physicists who has gone on to focus on complex networks, albeit these days in neurons and not people. I'm not entirely sure how this biases me, but it does mean that I've thought about a lot of the issues under discussion in this article.

A network doesn't need to be 100% right, it just needs to be useful. This is generally true for all of science, actually. There are indeed far too many ways too people can be connected for a monolithic social graph to be remotely useful. Sure, you can weight edges, annotate edges, whatever, but at the end of the day this introduces arbitrary amounts of complexity. Even in a situation where the nodes are clearly defined, such as people, the network you want to think about depends hugely on the use case one wants from the network. If one is interested in the spread of trends, then it's quite possible that the network isn't entirely based on social engagement, but is based on those who you actually see with regularity, including random people on your college campus you've never spoken a word to. Or if it's disease spread, it can be people who use the same doorknob as you. At the end of the day, the edges are the things that make two individuals "close," whatever close means in that context. Different uses will have radically different definitions of edges, as is very well mentioned here.

My feeling from papers I've read and presentations I've seen is that the academic researchers who study networks of people understand this. They know this because it's damned hard work getting good datasets, and the limitations of that data come to the fore when you work with it. I wouldn't be at all surprised if companies don't often understand the weaknesses (or the strengths!) of social media data, though.
posted by Schismatic at 10:57 AM on November 9, 2011 [5 favorites]

I don't think many things are representable by a pure, unlabled graph. The road system is probably one exampe. You can ask questions like 'can I drive from x to y'. I thin the author is right in that you can't really do much with a social graph once you have one, other then maybe try to market stuff to popular people.

The other thing though is that for the edge lables, rather then having people sit down and figure stuff out and lable everything with defined lables, you can use machine learning on their interactions. Even doing (at full creepyness levels) doing natural language preoccessing on their on-site conversations. An edge doesn't need to be a fixed term, but rather parameters for a statistical model.

Actuall I remember reading about zuckerburg, back in the early days of facebook writing code to predict whether or not two people would start dating (I.e. set ther relationshp status). Supposedly it was pretty accurare

(I'm not saying companies SHOULD do this, jsut that its possible )
posted by delmoi at 10:58 AM on November 9, 2011

my favorite guy on the internet whose name I can't pronounce

MATCH-ey tseg-WOV-ski
posted by Meatbomb at 11:24 AM on November 9, 2011 [1 favorite]

MATCH-ey tseg-WOV-ski (ignore the presenter's attempt the surname)
posted by Busy Old Fool at 11:33 AM on November 9, 2011

a common problem with tech writing which the need for people to proclaim "everything you know about topic X is wrong" its really annoying

Stealing your shtick, huh?
posted by yerfatma at 11:45 AM on November 9, 2011 [3 favorites]

I stumbled across Idlewords years ago, this guy is great. I'd love to meet him someday.
posted by roboton666 at 11:50 AM on November 9, 2011

Asking computer nerds to design social software is a little bit like hiring a Mormon bartender.

Actually I live in Utah and have been served by a number of Mormon bartenders, and most of them are quite competent. Mormon waiters/waitresses, on the other hand, should not be trusted to bring you anything as complicated as an Iced Tea.
posted by mmoncur at 11:54 AM on November 9, 2011 [1 favorite]

This was an interesting read.

I did my Ph.D. in archaeology as a social network analysis of site size and location - applying a network implementation of a location-allocation model. The interesting thing is, it found a bottom-up pattern of least effort optimization across a large region that could never have been consciously recognized by the actors on the network (who, in their tens of thousands across two millennia, were only known by the congealed material residue of their behaviour).

Basically, it found that some of the apparent settlement complexity was an emergent phenomenon of long term social interaction mediated through material culture.

What's my point? I don't know, exactly, but my finding was that not all, but some, human behaviour was the unintentional byproduct of the social network itself. As archaeologists, we look through the wrong end of the telescope, a lot. But those saying that social network analysis is incomplete, or flawed, or doesn't represent their own intentionality, agency, or social constructs, may well be missing the forest for the trees. "We make our tools, and thereafter, our tools make us" (Marshall McLuhan).

On another topic, related slightly, it is absolutely astonishing to see how the Indigenous Peoples of B.C., a diaspora of small communites and urban aboriginals, coalesce around facebook. The "Proud to be Kwakwaka'wakw" facebook page is one of the most vibrant and interesting examples of how creating a venue for social interaction then becomes itself an agent of social interaction.

Or, shall I say, "was" because it is one of the many groups that facebook has been killing off or stifling or crippling by forcing to a new format and who knows if that will happen and whether the 817 members will migrate to the new one. Therein lies the ephemeral nature of the material means of connection that is the internet.

The Native American photos facebook pages have also spawned a real virtual community focuses on the circulation and curation of images.
posted by Rumple at 11:59 AM on November 9, 2011 [8 favorites]

What I Didn’t Write About When I Wrote About Quitting Facebook
posted by mrgrimm at 12:19 PM on November 9, 2011

I saw the statistic recently that 98% of Americans are involved with social media. (And how we laughed at that statistic)

Here's to the 2%: 1.85% are unable to partake due to being homeless, the other 0.15% sit in awe at the stupidity of trusting a Corporation who makes money from unique access to your life, social circle, wants & desires.

Total fruitcake land. Guess no-one ever taught scientia potentia est in America.

And no, you're not an unique snowflake, you're a product. Your online profile, if you've been in any way honest & open is a fucking resource for other people to make money from. You've whored yourself out without asking for a penny back, other than that "warm fuzzy feeling" when someone you've never spoken to, and never will meet, "friends you".

Here's a tip: Emotional Connection is a marketing tool for these people.

Baaaaaaaa.
posted by Cheradine Zakalwe at 12:46 PM on November 9, 2011 [1 favorite]

Hm. I think I'd rather be a sheep than a jackass.
posted by Zozo at 1:03 PM on November 9, 2011 [4 favorites]

> Here's to the 2%: 1.85% are unable to partake due to being homeless, the other 0.15% sit in awe at the
> stupidity of trusting a Corporation who makes money from unique access to your life, social circle, wants & desires.

Speaking as someone who didn't join AOL then and has not joined Facebook or Google+ now, I'll drink to whatever slice of the 0.15% is foreveralone pariahs. Just don't any of you show up and want to drink with me, unless you BYOB.


> Hm. I think I'd rather be a sheep than a jackass.

Let us bray.
posted by jfuller at 1:10 PM on November 9, 2011

And no, you're not an unique snowflake, you're a product.

...which is already enough of a cliché to be the title of this thread. I think you're misreading the room if you think that and a sheeple reference are going to be persuasive here. Moreover, we've had a number of conversations before about social networks and privacy in general, whereas this discussion is about a slightly more nuanced argument regarding the possibility and utility of social graphs.

Oh, and I've never had a Facebook account - this is not about me being offended by oh-so-radical criticism of Zuckerberg's empire, this is about trying to have a decent, respectful conversation on Metafilter.
posted by Busy Old Fool at 1:11 PM on November 9, 2011 [3 favorites]

Here's a tip: "wake up, sheeple", as rhetoric, has never achieved anything.
posted by murphy slaw at 1:11 PM on November 9, 2011 [1 favorite]

Cheradine Zakalwe, this is sort of an example of that "coming on strong" stuff you and I have talked about. Please cool it.
posted by cortex at 1:38 PM on November 9, 2011

Cheradine Zakalwe: "Baaaaaaaa."

Man, it's an odd thing you do - going from all fiery and erudite to slavering caveman within the space of a sentence and back over and over again.
posted by vanar sena at 1:39 PM on November 9, 2011

> Baaaaaaaa.

Are you familiar with the website you're currently reading?
posted by ardgedee at 1:46 PM on November 9, 2011

1.85% are unable to partake due to being homeless

Got a local library? Ever been to it?

Also this:

"Facebook, Twitter and blogs are being used by homeless people to get help and stay connected. They use these social networking tools to find housing and food banks nearby."

Or this: "Each day, we distribute enough food to provide over 93,000 meals to people in need throughout San Francisco and Marin. You can help - make a gift here: http://bit.ly/GiveFood"

Why social media for homeless people? Why not?

Your online profile, if you've been in any way honest & open is a fucking resource for other people to make money from.

If you are attractive enough and have good photos, your online profile can also be a good resource for other people to fuck to. *animated winky face*
posted by mrgrimm at 3:42 PM on November 9, 2011

Social networks exist to sell you crap.

Also, I don't think this is true.

*Commercial* social networks exist to monetize your information. Advertisers and merchants use purchased information to sell you crap. Your information is used to attract new users and encourage existing users to post more and click more.

*Commercial* social networks also exist to sell advertising. They do have a vested interested in encouraging positive ROI on advertising, but they don't inherently care whether or not you buy crap. They care whether or not the people who want you to buy crap buy advertising.

I don't think Facebook or Friendster or Myspace were created with the idea of selling people crap. I think they all figured they'd get the critical mass audience and monetize one way or another ... probably by selling crap.

There are also online social networks that aren't commercial, obviously.
posted by mrgrimm at 3:47 PM on November 9, 2011

I must disagree with his sentiment that "Asking computer nerds to design social software is a little bit like hiring a Mormon bartender" because imho the two most interesting social-ish sites are github (previously) and stackoverflow (previously), although perhaps that confirms his statement. <shrug>

I loved his ending however :

"The funny thing is, no one's really hiding the secret of how to make awesome online communities. Give people something cool to do and a way to talk to each other, moderate a little bit, and your job is done. ... MetaFilter, Reddit, LiveJournal and SA all started with a couple of buttons and a textfield and have produced some fascinating subcultures. And maybe the purest (!) example is 4chan, a Lord of the Flies community that invents all the stuff you end up sharing elsewhere: image macros, copypasta, rage comics, the lolrus. The data model for 4chan is three fields long - image, timestamp, text.

Now tell me one bit of original culture that's ever come out of Facebook.
posted by jeffburdges at 8:06 PM on November 9, 2011

heh, I was wondering how long it would take for this to surface on mefi.

The fact that is was roughly 5 hours reconfirms my respect for this place.

Yes, Yes, Yes: make something that people like.
posted by grubby at 10:14 PM on November 9, 2011

otherwise put: "Occupy Social Networks"
posted by grubby at 10:18 PM on November 9, 2011

Maciej is turning into my favorite guy on the internet whose name I can't pronounce.

Macy-J. I will hear no other answer.
posted by cereselle at 9:48 AM on November 10, 2011