UCCSS (University of California Computational Social Science): Hilbert Intro2 Examples

UCCSS (University of California Computational Social Science): Hilbert Intro2 Examples

October 22, 2019 0 By Stanley Isaacs


So what are some of the examples of computational
social science and this, following, this last part will basically be a short overview of
the content of the course because that’s what we will do in this course, this introductory
course is to give you an overview of different examples of doing computational social science
and we will expose you to some of the methodology involved. And we will start with our computational scientific
methods and you will see this framework a lot. This framework will guide us through the entire
course and we will start here in the upper left corner with the empirical work which
in computational social science is big data. Often as I already mentioned the digital footprint
that we have-so the digital footprint is really a footprint you can think about it this way:
the digital footprint that we leave behind for example here with Twitter. And we just take this kind of data and then
with that, try to add to social science questions. For example, this here what you can see in
this video is when someone in Twitter says “good morning” or “buenos dias,” “guten
morning,” “bonjourno” and we see whenever we have the digital footprint there we can
also then see when people get up and you can find some interesting things from there. Nobody had a survey, nobody had to do an experiment,
nobody to ask anything to anybody. People just you know get up and say good morning
and now we can see some differences in when people do get up. This year using also Twitter data in Manhattan
in New York and we can see where people from different ethnic backgrounds live. So they just see what language do people write
on Twitter and you can identify some Russian communities, some Korean, some Japanese communities. Again, nobody did a survey, nobody asked anything
but now very clearly actually by block, we can identify where people from different ethnic
origins live because that’s the language they also tweet it. So they basically kind of like mapping out
the world around you with this digital footprint. Now this time you’re mapping out the world,
you’re not mapping it out geographically but socially so you’re doing social science
the advance of that a lot so in the past, we made maps that kind of look like this. That’s Monrovia, capital city of Africa
and for a few hundred years we mapped out geography of these cities and we knew that
in Africa, that’s what it actually looks like and now the same time especially in Africa,
only half of the people have a birth certificate so actually we don’t know how many people
there are, we don’t really have any register, we don’t have any solid empirical evidence
to do social science. The government doesn’t know how many people
actually live in the country so mapping out this area, socially it’s kind of like the
first mapmakers when they would write about “there be dragons”, I mean, I don’t
know what’s behind that but so this map was done in August 2014, only a few months
later in November 2014 that’s what that map looked like. Again we go from the “there be dragons”
here to that and they could fill out this map knowing actually where people are. How did they do that what do you think? They used cell phones. They use the digital footprint that people
leave behind with their cell phone and almost everybody even in Africa nowadays has a cell
phone so we don’t know how many people there are but if a cell phone walks around the street,
we assume that a person is attached to it. Actually that’s the best evidence that there
are people and how many there are so in this sense, we’re measuring out the world again
with this digital footprint. We can ask a lot of questions, social science
questions that before were either too costly to ask then or also to study. I already talked about the problems of relationships. For example, we could look at the digital
footprint of what are people worried about in their relationships basically what they
search on the Internet. So if we look at that birds go together with
the birds like your marriage or relationship, we see the word sexless marriage is by far
the biggest concern. Whereas in relationships, that’s not the
biggest problem. In relationships, sexless relationships is
the second concern that people have and the other comparisons for example, if you see
what husbands and wives search on the internet we can see that the search for my husband
won’t have sex with me and my wife won’t have sex with me that’s kind of like equal
right here both about 1,000 so that’s kind of like an equal concern among husbands and
wives. However now if you go into relationships we
can see that my boyfriend won’t have sex with me is twice as big of a concern that
my girlfriend won’t have sex with me so yeah I don’t know what’s up with all these
boyfriends right? Now so this is again a digital footprint. It might be a sampling issue, maybe girlfriends
search them on the Internet and boyfriends not. So there’s concerns that we will have to
talk about and when we talk about it when we do computational social science. But there is a lot more evidence out there. The founder of this dating website called
Ok Cupid also wrote an interesting book and he basically studied what people are doing
on this dating website and one thing he studied is what are the ages of example woman and
what age of profiles of man do they look at. So here he mapped it out so we have the ages
of woman and we can see that women in their 20’s for example with 24 they look for guys
who are a little bit older. A 24 year old woman would look for a guy who
was like 25, 26 where as the ladies in their 40’s for example with 46, 47 they would
look for younger guys. You know, in their late 30’s so that’s
why the diagonal is kind of like cross. The younger ones look for the older ones and
the older ones look for a bit younger ones. Very interesting. It’s just as the digital footprint they
didn’t produce it really they just looked for that and now we can have evidence of it
by studying this digital footprint. What do you think this graph looks like for
men? Yeah, right? That’s really what we found. That’s really what it looks like. Independent from their age, men always look
at profiles of women who are in their early 20s. You know traditionally that’s kind of like
been a joke it’s been a kind of like running gag. So for example, you would say it’s like
a party joke. “Hey dude you know women say we men are
not consistent. We’re very consistent. We always like women in their 20s heh heh
heh.” You know it’s a joke and we laugh about
it but now we have empirical evidence about it and having this empirical evidence actually
makes you think like being a social science researcher. You know, living in a society with a divorce
rate of 50%. Having a big proportion of children growing
up with only one of their biological parents. Building societies like that you start to
think like what kind of species are we? And so here we just collected, these people
did not know they were observed and it’s a very big footprint of millions of people
and that’s just naturally what they do so now we have the tools to ask these questions
to go deeper into trying to understand what’s going on with this kind of weird thing that
we call humans and what kind of species are we and try to figure that out in order to
see how we can make all of society better. And given all of these opportunities, people
say that what this big data digital footprint is, is kind of like equivalent to what the
telescope was for astronomy. So before the Inkas and ancient cultures already
looked at the stars and mapped out an amazing amount of stuff but with a telescope, we converted
astronomy into a science. We could finally see and we could see very
far with this kind of telescope and they say you know that digital footprint allows us
to finally see society with that level or what the microscope was for biology, we always
had an idea of the kind of like that’s how cells were actually work and behave. But the microscope allowed us to see them
and with that biology became a really strong science that allowed us to make predictions
because we could not understand and the digital footprint, it’s kind of like the equivalent
of what the telescope was for astronomy and the microscope for biology. And during this course, we will work with
webscaping tools, so basically you have this digital footprint here on your favorite Internet
site and we develop some webscraping tools that then that a machine helps us to derive
data from that. And with this data, we can make analysis so
we will develop some webscrapers together. You will develop a web scraper yourself and
scrape some data in order then to have some empirical observations about society and then
to do some analysis which brings us. Our second part of our computational social
scientific methods: analysis. And we start with social network analysis
just because well society is actually a network, a network of people and it’s an extremely
powerful tool enabled by the digital footprint and by computational techniques. So as I already said, the social is actually
a network and we have networks all around us. Not only what we now call a social network
that is kind of like an online social network, there can also be an “offline” social network. And we found for example, James Fowler and
colleagues at the University of California – San Diego, found that even without digital
networks, traditional networks, things like happiness are kind of like contagious. So even if out of a second or third degree,
out of a friend of a friend of yours, is happy, that can kind of like can an affect on your
level of happiness as well. So the social networks are extremely important. Entire nations can basically be thought about
as networks. There is an entire cluster of people that
basically map out the networks of industry leaders and government leaders and in which
__ and which board of directors they sit and you won’t be surprised if I tell you that
you don’t need to map many of these people because it’s the same people sitting in industries,
sitting in the government, sitting in the committees, and actually what our nation is
and how our nation is run, you can very quickly map out by looking at this social network
that constitutes our society. And I won’t go deeper into that; we have two
lectures on social network analysis. One of my mentors, Manuel Castells, wrote
a thousand-page trilogy on the rise of the network society and you can, well, the social
is basically the the network; you can explain the social phenomenas in terms of these kinds
of networks. Now the computational approach and the digitial
footprint has been very important in order to reveal these networks; as I said, these
networks always existed and they have important effects on us but the digital footprint allowed
us to reveal these connections and to make a sign of a more former science out of it. Computational tools are also very important. Let me show you an example of how important
it is if you do social sciences. Traditionally, if you have a social science
problem, for example, here we have a bunch of people, we have 8 people, with different
characteristics traditionally what we would do is look at who they are. Okay so, who are these people? Okay aged people, some them have education,
you know, that’s why they have that educational hat on. Some of them have a computer so they do computational
techniques. And oh, 2 of them have red pants. Okay, so there are some characteristics and
now we can make a theory out of that. Let’s see if we can find some relationship
and see who of these guys do computational techniques and who do not. So traditionally, we would look at the characteristics
so with education, without education, with computer, without computer, and then we just
count, put them in boxes. We put people in boxes, so 3 of them have
a computer and have education, 1 of them has a computer but doesn’t have an education,
1 of them has an education but no computer, and 3 of them don’t have a computer an don’t
have an education. Then we would do our famous analysis technique,
for example, we square, we run a correlation, a regression and we can see… yeah, education
has something to do with the use of computational tools. Fantastic! Policy recommendation; therefore, would be
well. Look at the people who have education… Well, not so fast… maybe the people with the red pants also have
something to do with it. Okay so let’s check that. Let’s check with the red pants just to make
sure we’re not on the wrong track. We have 2 people with red pants, one with
a computer and one without, and 3 people without red pants and with a computer, and 3 people
without red pants and without a computer. So there’s no relationship here between, right? There’s just as many people with red pants
and with computers, so okay, we proved it: red pants have nothing to do with it. Don’t worry about the red; what should red
pants have to do with it, right? So we do our policy, good. As I said: focus on the people with education
to push computational methods. Now, as an example, it turns out that network
structures often are very important because especially something like innovations, computational
methods here being innovations, they spread through social networks. So if we now would reveal, in this hypothetical
example, this social network among these people, these people, are not all independent, they
kind of like hang together and it’s not like everybody is connected with everybody you
know, and we often do find something like this here: so we have kind of like one cluster
on the one side and one cluster on the other side and in this hypothetical example, the
people with the red pants DO matter. If I want the innovation of computers spread
from one side to the other, I actually have to build a bridge between the people with
the red pants but who are the people with the red pants? Well, they are innovators, they are agents
of change, they are so innovative that they are wearing red pants. So actually the best way to foster the spread
of an innovation, if you look at it from this perspective, that their social networks would
NOT be to do something to help people with education and focus on that, NO! You can do a much cheaper intervention by
focusing on these kinds of innovators. How do you detect them and in this case, I
reveal the social network behind it. Now this is just a hypothetical example but
that’s often what we find and what the digital footprint helped us is to get this additional
piece of information: who hangs together with whom. Before, it was like a you know, ______ You
knew, but nobody else knew who you were actually hanging out with and now we can actually leave
that behind and we can analyze it. And society is as much about who you are as
about with whom you are, as kind of like your mother told you, right? “Be careful with whom you hangout with, because
who you hangout with…” And it’s true! It’s not only about who you are but tracking
social networks, it’s only became actually viable in a massive approach thanks to the
digital footprint. And now we can analyze these social networks. Now, social networks have very intricate structures. Just look at this network here; you can see
and analyze these structures, and we will do that throughout 2 lectures. In one lecture, we will look at the structure
of the network and in the other lecture, we will look at how networks evolve, how they
actually change over time and we will understand more about that. Both of that is very relevant for doing social
science because as I already said, you can get a lot of mileage out. For example you can make policy recommendations
to make the world a better place, you can get them much cheaper. Here is an example that used computational
social science that controlled the spread of a disease in schools. Instead of spreading innovations, I know try
to stop the spread of a disease that also goes through social networks and we try to
contain it. So they found out that it’s not the red pants,
it’s children. Children play a very important role in the
community and the spread of a disease. So the main policy goal in order to try and
prevent the spread of a disease is, for example, well we have children, they are in schools,
so people, policymakers, would just close down schools. For example, that has a very high political
and social cost so if you think close down an entire school, parents have to stay at
home so they cannot work for the economy… that’s a huge hit for the economy if you close
down a school. Alright so closing down one school is up to
one hundred thousand dollars but just from the money that the parents lose, the economy
does not boosted with that. So let’s look for some smarter ways of how
we can control of a disease. Well if you would know the network structure,
we can much better see like, well, how can we stop it, right? So computational social science solutions,
for example mapped out here the network, and that’s what a network looks like in a school,
there are different clusters in that network. Well, no surprise, these are the school classes
and there are different school classes at different grades and between the grades, for
example, among all 2nd grade, or they also have a lot of contact but then in different
grades, there’s not so much contact. And basically what they showed with a lot
of computer simulations, and we will do computer simulations on social networks and study that
in detail, what they found actually: if you close down just one class, a class where there
are two infected students, that is just as effective, almsot as effective, as closing
down the school. Now you get 70% of the effect. If you close down the entire grade, let’s
say they are in one class, 2 second graders that are ill, you close down all second grade,
that’s still much less than closing down the entire school, you get almost as good of an
effect. That’s what this graph here shows. So how do we get all this social mileage,
social policy, well look at the network structure and analze and simulate these kinds of networks. Try to understand it and thanks to the digital
footprint and to computer simulations, we have these modern insights and can help wel,
make the world a better place basically. one very important analytical technique in
computational social science has to talk about is the basically artificial intelligence ever
talk about machine learning in about natural language processing the idea behind machine
learning let me say well we have so much data nowadays the date is so big big Jada that
this information processing here is helplessly overwhelmed with processing it Fight Fire
with Fire so the district brought us information overload it’s also used H2 technology in order
to make sense of it all right if I use machines to learn patterns in this information over. For example books if you read one book per
day for your entire life every day you were born book for 80 years you would read about
30,000 books and hopefully you’ll be able to process 12 in this in the in this process
of here is 160 million like you there’s no way shape or form process that so they are
all there and actually was mushy you can look for patterns you can look for what they are
and what kind of patterns and you parents maybe that we ReDiscover of taking you know
looking at a hundred and thirty million books not only is it find me can look at work take
a pattern different many different data or data important machines can do that all right
to machines look for patterns discover discover summer of these benefits and then what can
they do with it while they can do a lot with it actually modern approaches to artificial
intelligence only became viable with this big data footprint that you have artificial
intelligence has been around for a long time but things to this digital footprint that
we have things to the massive amount of data now machines can learn discover parents into
really amazing things so check out some of the videos I have here for you this for example
of pure is the Amazon inventory Warehouse so if you click something on an online retailer
like Amazon and want to buy something immediately one of these Bots artificial intelligence
but gets going and get your purchase this over here in the corner that’s what I modern
car manufacturing company looks like what there’s no person involved here anymore it’s
machines for learned parent of how to construct a car it’s hundreds of robots working together
and if you find a human it’s probably some kind of it nerd to adjust the robots but cars
nowadays A basically build by artificial intelligence this down here is an example that I want to
show you his how quickly these machines can learn so it is an example of an artificial
intelligence from Google replace an Atari game a very old Atari game I used to play
that as a kid and the computer that he doesn’t know anything about the game it just knows
it’s kind of like a ping pong ball and it has to collect points so that’s that’s the
only two things you tell it it doesn’t know how the game works and that’s how quickly
it learns how to master the game check it out well that’s pretty impressive everything
became Innovative it actually innovated it figured out that putting the ball on top rated
everything I found I found a solution shortcut I play this game as a kid and it probably
for more than a few hours and I didn’t figure that one out to figure that out but honestly
I want me for sure found a very creative solution for that have to find so if we let machine-learning
lose on this data is find some kind of patterns that this process a year was not even aware
of that existed sometime this all to be looking at this black box of machine learning and
you don’t even still don’t know how to looking at it what it actually did and why it is so
much better than humans it doing what it does and it became better than human in many many
areas for example in Radiology the discovery cancer cells nowadays artificial intelligence
what’s the self-driving cars in Beachwood go watch it so it’s it’s it’s more secure
even then people we can go on and on artificial intelligence as being as everyday with some
stupid looking through these massive amounts of data and discovering discovering patterns
creating knowledge and that’s what communication Sciences about right so machine learning can
help us with that additionally latest to the point what are we actually doing when we when
we create new knowledge what are we doing when we do science what is intelligence working
with artificial intelligence we also happen to understand much more but this entire idea
is about look at our Technologies and then actually understand what something is about
for example look at flight human flight we we always start something flying and has to
do with feathers the only thing we saw flying in human history rubber it’s better than any
to slap them probably be able to fly it had no idea would ever dynamic the brothers right
not too long ago a hundred and something years ago the first time they were free 30 m 400
ft Elvis killed himself said no idea how this stuff works so they built the first flying
machines why dangerous without knowing what they were doing it survived many others many
other things gambling at that point but then was the IDS flying machine flying horse fly
actually is supplying but it doesn’t have to do with feathers that has to do with the
curvature of the wings so he sees huge airplanes of levees little tiny wings but the recruiter
actually researching their socket what does extra butter stain is about right and we could
build all kind of different flying machines that flying machines that nature never came
up with helicopters fighter jet planes and Rockets and after the brother tried only 60
years later 60 years later we were flying to the Moon nature never invented something
that could fly to the moon at least not on no Evolution never came up with that but once
you understand the principle of what it actually is we were doing things that nature with his
evolutionary tinkering never came up with them understood all kind of flying techniques
and Internet Protocol something similar is happening right now with intelligence came
up with one solution Asians problem kind of like the birds and the feathers now seeing
how much it’s restoring level Plains of the bigger picture word actually is to create
knowledge what it is to learn by these machines that we created actually do exactly that and
his machines get so good at is she becomes a little bit scary discussion if you open
a newspaper nowadays you see it you know that we don’t have to go to the Terminators and
to you know 222 how artificial intelligence will let you know because we had a serious
pieces with no need to go there very many people are buried in artificial intelligence
is taking the jobs for example and even you know the biggest company of Lititz companies
that use attitude much to machine learning artificial intelligence massively they have
a very healthy dose of respect to water open up the artificial intelligence if you want
to do artificial intelligence companies like Google and Amazon they open up while they
open up in an empty brain NPR official brake they don’t tell you how they trained it
training the sprain so they have the date of the traded risk I like they have at University
graduate there as well newborn baby you can train it yourself and it’s available play
Fever play the some of the artificial intelligence that’s open up by companies that buy Bitcoin
they do artificial intelligence the movie able to play play with it and in order to
do some sound machine learning for our cells we have a lap call braids with intelligence
traditionally VR with afro new technology because it might be taking our job at jobs
it might replace us and that’s a very old fear old story in or the 18th in the 1800
famously destroyed cotton and weaving machine because they were taking the jobs they sold
like oh my goodness what is cotton and weaving machine won’t be work anymore we still have
work because Tonya Scott MPV machines we automated that and then many tasks and intelligent kind
of like automating Whitney Houston machines and that’s something if you go to one of the
founders of computers of computer science John for Newman and he famously said the best
we can do is to divide all processes into those things which can be done better by machines
and those which can be done better by humans and then invent messes chew so that was the
idea since the beginning of computer science and yesterday’s message to corporate collaborate
with our machines interesting enough that’s also good we’ve been finding when we try to
to measure ourselves through the machine human organ machine man against machine right think
about Chess so in the late 1990s the end of the century we lost that battle in chest took
our best chess player Garry Kasparov we went against the computer from IBM deep blue and
we lost that one alright so afterwards Kasparov you know you had a couple of choices of what
you do he could have gone home studying more chest or he could have gone home and bury
himself in the hiding in the woods since I owe my goodness the machines at the powerful
even work out what he did is he started to corporate Business Machines then something
called freestyle chest actually you know computers are allowed and you play chess with me this
what they found dear is that the most successful teams were not the Grandmasters interesting
Vietnam also not the supercomputers and The Grandmaster teams that are very good Incorporated
by this machine so they say one it guy and one mediator chess player but they were very
good bodies and they really knew what your juice and Incorporated with the machines very
well and they were beating the Grandmasters with a supercomputer the best weekend machine
learning separate these processes what can machines do better in discovering knowledge
and what can we do better and then call braiding retirement me find out that this is the most
powerful approach all to do science and in that sense to computational science computational
social science especially And that brings us to our last leg of computational
scientific methods and that’s theory. So we can also do theory with the help of
computers–we often do that with simulation so we simulate theoretical scenarios that
do not necessarily exist in empirical reality. For example, here I show you a simulation
from the United States Army simulating for example an occupational strategy in different
in different countries and how many soldiers are needed when in order to provide security. here you can see a chemical attack in Los
Angeles. There’s never been a chemical attack in Los
Angeles, but we can simulate it and just see, well, what would happen: how would people
move? How would people run? and here we simulate
traffic in Chicago which is calibrated by cell phone data for example, and we can see
that these little hills is where there are traffic jams and you can see and then we can
simulate what would happen if you would change the traffic lights–if you change the traffic
lights this way and that way and nobody has to suffer because we just theoretically explore
what would happen and once we find a good solution to reduce traffic jams then we would
implement it right, so we do computer simulations of something, of some scenarios, that never
really existed. Why is that so important and why is that so
important for the social science especially? Well, if we only go by data, by empirical
approaches, data can only tell us what are already happened in the past because data
per definition is always from the past. I mean the minute you record it the moment
already past and then you have data even if it’s real time data you know it’s still it’s
always from the past. Now, you analyze the past… what extrapolations
can you do? So for example imagine you analyzed the past
and you can see here this kind of tendency. How do you think this tendency will continue..
and there is nothing nothing in the data that allows you to actually to actually say otherwise
the data statistics forces you to continue to extrapolate it because only in the data
there’s nothing different but in theory it might continue like this right? And maybe there’s something or we might want
to change it. So either there’s kind of like a phase transition
or just changes the stationarity of the series is broken–that would be the technical term
or we ourselves intervene because we want to make the world a better place. Now we don’t have data about a world without
poverty without pollution and we don’t have data about that but in theory we can think
about how to do that– right and with that we can extrapolate these kind of like phase
changes even and that’s in the social science very important because more is often different
and we often find these non-linear phase transitions in social sciences. it’s easier if you think about it in terms
for example of physics we find them in all kinds of sciences so in physics for example
if you have water right you can heat up water and heat up water and heat up water—its
still water—but then suddenly its at one hundred degree Celsius more Celsius make suddenly
has a different result it becomes a vapor. And then you have the water and you cool it
down, cool it down, and still water, its still liquid, and then it’s zero degrees Celsius
more cooling suddenly has a different effect. it becomes ice right, and so there are these
phase transitions and we see them in society a lot. if you do public policy these are the ones
you want to detect because you can like invest, you invest, and invest and you want to see
like, when the system and its tipping points, where you can actually flip the system and
just with statistical extrapolation you cannot really get to it once you didn’t observe them
and often you have NEVER observed them even in your digital footprint they are not there
and social society is riddled with these kind of nonlinear phase transitions. Schumpeter, an important economist—the Economist
of innovation famously said that the history of capitalism for example is studded with
violent bursts and catastrophes. We come to the conclusion that evolution is
a disturbance of existing structures and more like a series of explosions than a gentle,
though incessant, transformation. We have you know revolutions basically change
society and it’s not like a continuous thing it started with bursts and explosions with
revolutions—be it technological revolutions, innovations and so forth and we have to understand
them and just extrapolating of what has been in the past won’t allow you to detect the
impact of a new technology for example and when you often allow to detect Well many things
that happen why is that especially in social science is because these non-linearities where
do they come from? well often they come from interdependencies
among us they are not linear but in societies we are highly interdependent and therefore
we influence each other and this leads to these kind of non-linear behaviors. Let me show you just to give you a little
intuition about that– So here you have a growth model—a model of growth and we have
these bunnies–bunnies of course grow. So we have two bunnies and they both reproduce
and they always double So one body doubles to know each one as a duplicate now we have
two bunnies on each side, right and the two duplicate So now we have four and now the
four duplicate OK so we get eight on each side. if we summed that up now, it’s a societal
few, here we have two families of bunnies, lets take a look at it from the society. we put these two families of bunnies together
right—the the same thing should happen: OK so a linear approach to that would be:
we have two bunnies and then we go to four, then we go to eight and then we go to sixteen
so if you extrapolate that forward just by summing up individual summing up families
or to get society a linear extrapolation that’s what it would give us. Now what we often find is nonlinearities—Why
is that? Because their interactions among us. So for example there are some carrying capacity. we cannot have– imagine we have this in there
that we cannot have more than four bunnies on this piece of grass just because there
is not enough grass, so there is interaction now–competition between the bunnies and what’s
the effect? Well if you go by our first model, we have
two bunnies—one one on each patch of grass and the two bunnies together on one pitch
of grass and then they reproduce so we have now, we have two and two and four on the other
side what happens now? What happens if both of these sides reproduce
once more? Well on the one side, just looking at it in
a separate they can still reproduce because I said the carrying capacity is at four, four
a patch of grass can nurture four bunnies but on the other hand we already ran out of
resources. So there is this kind of non-linearity that
we often find. Where does it come from? Well because we affect each other so more
is different. We cannot just extrapolate data and if we
have more people on planet Earth, and if a new technology comes, and if you continue
in this tendency…–it won’t go on like this forever. We have to then also see theoretically what
we can…what we can do, in order to solve it. And we do this with these kind of computer
simulations–it’s kind of like playing SimCity. Do you know SimCity? Well yeah that’s what it looked like when
I was a kid playing with it, a teenager playing with it…that’s what it looks like today. I know really unfair, really frustrating. And What I did here with SimCity—SimCity.edu
that’s actually a software they use in high school nowadays in order to teach children. Here that’s a simulation of how… what…studying
the effect of sustainable alternative energy sources on the employment of a city. So here we have a power plant, a coal power
plant, and I just bulldozed that. I don’t want coal power plants anymore, I
change the course of history, I bulldoze these corporate coal power plants and then see what
happens. I want to create some other… a small wind
power plant. OK—alternative energy, I put a wind power
plant there—lets see what happens. Whoa, the city is dangerously low in power
right now OK let’s put a large sort of power plant next to it and I would make the transition
and I can study society. Now this society does not exist in reality,
I just made it up. There, there’s no city like this that has
only clean energy–Actually I made it up, and I can see now what happens. OK so if you ever play SimCity you see these
amazing things like Oh for example, wow, people start to to gather and protest in front of
City Hall–interesting just like policy makers. I’ve been working in the United Nations for
fifteen years and that’s policymakers–they are as surprised as you and me about like,
whoa unemployment went up, Oh there’s a social protest… it’s not like you know, not just
like it’s really surprising it’s a very complex system what if something goes well they always
said well that I intended that but having intentions in such a you know complex system
is really, really, difficult but with computational tools we can now explore kind of like the
space of possibilities and we don’t have to intervene in society, we don’t have to change,
we don’t, we can simulate digital twins of our societies and explore some things and
can kind of like weed out some alternatives that we do not want to pursue in practice. Now there’s a lot of work to do in that field
of these non-linear interactions that we can simulate and see things how they might go
forward even so we don’t have any empirical data to say in the words of Stanislaw Ulam
and Johnvon Neumann, which I already presented. Using a term like nonlinear science is like
referring to the body of Zoology as the study of non-elephants. Almost everything is non-linear. The only thing that’s linear is linear and
there’s only one way you can make a line it is infinite numbers of how you can not draw
a line a non-linear thing so there’s a lot of science to be done and if our generations…There
is a lot of low hanging fruits that we can pick–but for our generations especially,
with these computational tools it’s not easy, it’s not easy to be solved. The ream of possibilities is just too big
and we have to go forward exploring that. If you want to know more about this argument
I have an entire TED talk that I gave about this topic–about the limitations of working
only with data and why we want also to work with computer simulations. If you want please go ahead and check out
this TED Talk of mine. Oof, alright. Uh.. That’s been a lot and I hope you didn’t
watch all of that just in one sitting because there are too many things it involved that
I showed you here. Um… Let’s try to wrap it up, let’s try to
wrap it up and see and take a look again at our entire framework of what we’re doing
here in this course. I said well we start with empirical, going
on to analytical to the theoretical. And that’s bringing it all together that’s
actually the promise of computational Science in that case of computational social science
and there are big opportunities of bringing the empirical , theoretical, and analytical
together with these kind of new tools. and in the social sciences I already said
the social sciences traditionally people have question is that really a science or is more
art. and we have natural science, we do science right? Social? I don’t know, it’s more like you guys kind
of make it. But nowadays, actually the social science
has become maybe the only data complete science that we have, right? Other sciences don’t have as much data as
we have. historically we were running experiments on
yes, true on a collection of undergraduates who needed extra credit . And we were running
them through a lab, sixty at a time. And then we made grand conclusions on that’s
how society works. But nowadays, we have actually a tracker and
each one of us leaves a lot of information behind. And that’s very different from other sciences. So for example if you think about ecologist
they wanted to do that they wanted to actually do a computer simulation of all life on earth. Now, they don’t know where the fish are in
the ocean and they officially said that the biggest stumbling block is obtaining the data,
to parameterize and validate the model. And so they had automated cameras at the bottom
of ships and plankton recorders and they set up videos in the jungles to see when the puma
comes out. and to monitor where the fish is. But they don’t have the data of where all
life actually is. But we, as an evolved species leave that behind,
with our credit cards or our mobile phone or with our social networks and all 7 plus
billion people on planet Earth basically in one shape, form, or another leave this digital
footprint now behind. And that allows us to then to calibrate these
kind of computational models that I was showing you and basically study society. So there’s a lot of promise that social science
traditionally was one of the weaker sciences, now comes along during the next century, becomes
one of the sciences we make most advances and it is a very exciting field. We see physicists, biologists and so forth
flocking into the social sciences because of the big opportunities are there. I want to leave you with one last caveat which
is very important, especially when we do social science. Because in social science, we often also with
our results affect society. That’s different from other sciences. If you are Mr. Newton and you study how the
moon goes around the earth and the earth goes around the sun. Once you figure that out and then maybe as
a result of that, maybe you can fly to the moon a few hundred years later, but that didn’t
really change how the moon goes around the Earth. It didn’t change that scenario. In social science, once we figure something
out, we also intervene. And it’s very important to remember that all
models all science is wrong. And we will talk about this in this course
as well at the end of the course later. but but some of it is useful. Now its wrong, how can you imagine that because
we always have a snapshot of reality. So for example you do network analysis that
might be your network that you analyze. But the truth is it’s part of a bigger network
which is part of a even bigger network. But you might as well take in the same big
networks, so these two are the same, and and cut the pie a different way. Now you have two models of this reality, and
they lead to different conclusions. So the only model that you could really work
with, the only working model of Society would be Society itself. Well that would mean that you need to have
Society on your desk. You can’t even have Society in your computer. You have to abstract from it because you would
have to build a one-to-one model of society. You might as well use this. That doesn’t work. So all models are actually abstractions. We leave stuff out and that makes them all
wrong. And it can lead to two very important consequences. I want to leave you with one example, a recent
example of Alan Greenspan, the longest-serving chair of the Federal Reserve of the United
States, basically the person in charge of monetary policy in the country, the longest-serving
chair of the fed and when the economic crisis hit in 2008 he was called to Congress and
he was asked about in Congress what went wrong Mr. Greenspan? Why didn’t we see that coming? So many million people are suffering right
now around the world it is a big big echo. You could feel it actually in very poor countries
as well. People not only losing houses here in the
United States, people went hungry because of this economic crisis. So why didn’t we see that coming? Now check out this little clip here to understand
what Mr. Greenspan had to say in his defense. … Statement that you delivered the whole
intellectual edifice of modern risk management collapsed. you also said “those of us who have looked
to the self-interest of lending institutions to protect shareholders Equity, myself especially
or in the State of Shock and disbelief.” Now that sounds to me like you’re saying that
those who trusted the market to regulate itself, yourself included made a serious mistake. Well I think that’s true of some products,
but not all. I think that’s the reason why it’s important
to distinguish the size of this problem and its nature. And what I wanted too point out was that the
excluding credit default swaps, derivatives markets are working well. Well where do you think you made a mistake
then? I made a mistake in presuming that the self
interest of organizations specifically Banks and others with such that they were best capable
of protecting their own shareholders and their equities in the firms. And it’s been my experience having worked
both as a regulator for 18 years and similar qualities in the private sector especially
10 years at a major International bank, that the loan officers of those institutions knew
far more about the risks involved in the people to whom they lent money, then I saw even our
best Regulators at the Fed capable of doing. So the problem here is something which looked
to be a very solid edifice and indeed a critical Pillar To market competition and free markets
did breakdown and I think that shocked me I still do not fully understand why it happened
and obviously to the extent that I figure out where it happened and why I will change
my views. If the facts change, I will change. You had an ideology, you had a belief that
free competitive, and this is your statement . I do have an ideology, my judgement is a
free competitive markets are by far the unrivaled way to organize economies we’ve tried regulations,
none meaningfully worked. That was your quote. You have the authority to prevent irresponsible
lending practices that led to the subprime mortgage crisis you were advised to do so
by many others and now our whole economy is paying its price. Do you feel that your ideology pushed you
to make decisions that you wish you had not made? well remember that what an ideology is, is
a conceptual framework with the way people deal with reality. Everyone has one. You need an ideology to exist. The question is whether it is accurate or
not and what I’m saying to you is yes I found a flaw, I don’t know how significant or permanent
it is, but I’ve been very distressed by that fact. But if I may, can I just finish an answer
to the question. You found a flaw in the reality. A flaw in the model that I perceived as critical
functioning structure that defines how the world works so to speak. In other words, you found that your view of
the world, your ideology was not right it was not working. That’s precisely the reason I was shocked
because I’ve been going for 40 years or more with very considerable evidence that it was
working exceptionally well. So he basically said a few things I want to
point out. He said, well I made a mistake by presuming
that my model was actually wrong. He literally said I found a flaw in the model
that I perceived is the critical functioning structure that defines how the world works. So he had a model of how the world works and
that’s how we operate it, the biggest monetary pool in the world, the United states and then
he found out whoops, there was a flaw in my model, right? And that had big big implications for many
people. So when we do social science, its not only
exciting and very relevant, it also has consequences, And that leads us to our last point. If you have the wrong model, it can be that
you have very good intentions. I don’t think Mr. Greenspan had bad intentions. He might have very good intentions and that’s
how he testified in Congress right? I just want to do the best thing, but his
model was wrong. And that can be very counter intuitive because
you know usually in a comic world, it’s very simple to say you know usually good people
do good things and bad people do things. So they have the superheroes and the villains,
right? But in reality, it’s more like, you know,
there are some good people who do bad things and there are bad people who do good things
and that might be very counter-intuitive. But it happens all the time in social sciences. That’s why we also have to be aware of our
models and how important how influential the outcome are of doing social science and that
brings us to our last point of what we will have to talk about in this course and that
is research ethics. You really have to see if you do social science,
we have to first make sure that we follow the scientific principles, that we don’t do
harm when we do it. When we follow the scientific principles,
and that then as an outcome as well, you have to be very aware of what the outcome of our
conclusion. Because more often than not, it will affect
real people. Alright, that now really wraps it up. We kind of like rattled through three questions. First question is why computational social
science and why computational social science now. And I said well two things, first of all,
the technology is there, it’s all over society, great opportunities. Many low hanging fruits that we can explore
in order to understand society better. And second of all, social science we don’t
really understand it. We don’t understand society as well. That also leads to the fact that there are
many discoveries to be made and many Nobel Prizes to be won. And I hope you will join us in this pursuit. Second,what does computation social science
cover and I basically said well it affect all of the scientific method and we developed
this framework that we will actually go through in this course and we will come back repeatedly
to see the different aspects of traditional scientific methods and how we can do them
computationally. And last but not least, I went through some
of the examples of computational social science and the rest of the course, that’s what we
will do. We will go through some examples and Hands-On
exercises. So you will work with webscraping, you will
do social networks you will work with artificial intelligence with machine learning, and we
bring it all together and do computer simulations. You know computer simulations and then bring
it all together at the end webscrape, do social networks, run it through machine learning
and do a computer simulation and at the end of the course all of that will have been brought
together and we run through this cycle of a computational social scientific method together
and hopefully a few times so we get really comfortable with it and we get exposed to
the different ways we can do science, social science nowadays. Alright, I hope you enjoyed this first introduction
as much as I did