On Github mclevey / integ120
Dr. John McLeveyUniversity of Waterloojohn.mclevey@uwaterloo.cajohnmclevey.com | networkslab.org/Fall 2016, Knowledge IntegrationUniversity of Waterloo integ120.f2016.slack.com
Lecture Slides (Updated Continuously)
40/40/20: Research, Teaching, Service
Teaching Assistant
interdisciplinary systems thinking
leveraging diversity and social capital for good
getting to know the "big picture"
...
Systems Thinking Paper (Nov. 8, 30%)
Knowledge Networks Paper (Dec. 5, 30%)
10 Comprehension Quizzes (Ongoing, 30%)
Engagement / Participation (Ongoing, 10%)
You will co-author a paper that uses course readings to apply interdisciplinary systems-thinking to a contemporary problem. In addition to interpreting the problem from a systems perspective, you should compare possible interventions into the system and explain why some are more likely to succeed (i.e. make things better) than others.
You will co-author a "knowledge networks" paper on any disciplinary or interdisciplinary research area of your choosing. To write this paper, you will have to search a database like the Web of Science or PubMed and download meta-data on published journal articles. Then you will use (very simple) software such as VOSViewer or metaknowledge to interpret the high-level topics and network structure of your chosen research area. Then, you will select a small number of important papers (~6), explain why they are important, and identify their primary intellectual contributions.
I will drop your lowest grade.
integ120.f2016.slack.com
Laptops may be used in the classroom on the honors system. If I see Facebook, email, an IM client other than #slack, a newspaper story, a blog, or any other content not related to the class, I will remove 1 point from your participation grade on the spot. No exceptions.
You will need a laptop for all classes marked as "computing" in the syllabus. You will need a laptop, a tablet, or a phone for all classes where a quiz is scheduled.
VOSViewer and metaknowledge
Please complete Get to know people and help John make this a good class survey, which is active on LEARN until the coming Tuesday. This is to help us get to know one another, and to help me make this a better course.
I will also complete the survey and you can see my responses. I will discuss the data in class but I will not disclose names. I would like to make parts of your survey available to the rest of the class so that they can get to know you better, but I will only do so with your permission.Watts Preface & Ch. 1
Duncan Watts
Today: "Preface" & "The Myth of Common Sense"
(Photo Credit Michael Dolan) flickr
Commonsense is practical and can deal with everyday concrete situations on their own terms. It is essential to function in everyday life.
Common sense is "common" only to the extent that people share sufficiently similar social and culture experience. Collective tacit knowledge.
+ change over time, inconsistencies in our beliefs, etc.
Fragmented, inconsistent, and self-contradictory common sense doesn't pose a problem in everyday life because so many problems are small and independent of one another. It becomes a problem when we want to use common sense to solve problems that are not grounded in the immediate here and now of everyday life.
Examples?
"Bad things happen not because we forget to use our common sense, but rather because the incredible effectiveness of common sense in solving the problems of everyday life causes us to put more faith in it than it can bear."
Watts Ch. 2
Duncan Watts
Today: "Thinking About Thinking"
Duncan Watts
Today: "The Wisdom (and Madness) of Crowds"
"We claim to be saying that the Mona Lisa is the most famous painting in the world because it has attributes X, Y, and Z. But really what we're saying is that the Mona Lisa is famous because it is more like the Mona Lisa than anything else.
"We want to believe that X succeed because it had just the right attributes, but the only attributes we know about are the attributes that X possess; thus we conclude that these attributes must have been responsible for X's success."
How do you get from one level of reality to the next?
Many of the most interesting / important problems cross levels.
"We do not speak of the genome as if it behaves like a single gene, not do we speak of brains as if they behave like individual neurons, or ecosystems like individual creatures. That would be ridiculous. When it comes to social phenomenon, however, we do speak of "social actors" like families, markets, political parties, demographic segments, and nation-states as if they act in more or less the same way as the individuals that comprise them."
What does it mean to say that crowds behave differently than the individuals that make them up?
How does Granovetter’s riot model illustrate the argument?
Photo: Vancouver RiotCommunity, "Remedial Chaos Theory"
"If we were to imagine history being somehow 'rerun' many times, therefore, explanations in which intrinsic attributes were the only things that mattered would predict that the same outcome would pertain very time. By contrast, cumulative advantage would predict that even identical universes, starting out with the same set of people and objects and tastes, would nevertheless generate different cultural or marketplace winners."
"When individuals are influenced by what other people are doing, similar groups of people can end up behaving in very different ways." We shouldn't sidestep this problem by using the notion of a "representative agent."
Watts "Special People"
Duncan Watts
Today: "Special People"
The goal is to get a message to Trudeau, but you can only send the email to someone you know personally. Whoever receives your email can only forward it to someone they know personally.
What is a plausible path it might take, based on your reading about the small world problem. Remember, you can only control the first step.
What did Jane Jacobs get wrong?
Even though there are "stars" in social networks, people are not airports! There are limits in the number of friends a person can have, and it doesn't come close to the number of connections in large airports.
"when you want to get a message to a graduate student in Novosibirsk, Russia, you don't think about whom you know who has a lot of friends, or goes to a lot of parties ... You think about if you know any Russians. ... Maybe you know someone from Eastern Europe, or someone who has traveled to Eastern Europe, or has studied Russian, or who lives in a part of your city that is known for its Eastern European immigrants."
In spite of this evidence, we still think of special people who are disproportionately well connected.
What is the argument?
"Contagion -- the idea that information, and potentially influence -- can spread along network ties like an infectious disease -- is one of the most intriguing ideas in network science."
...
"Influencers may exist... but not the kind of influencers posited by the law of the few."
"The reason is simply that when influences is spread by some contagious process, the outcome depends far more on the overall structure of the network than on the properties of the individuals who trigger it. Just as forest fires require a conspiracy of wind, temperature, low humidity, and combustible fuel to rage out of control over large tracts of land, social epidemics require just the right conditions to be satisfied by the network of influence."
The most important condition is actually a critical mass of easily influenced people. Not the presence of a few highly influential people.
Design? Take home?
Duncan Watts
Today: "History (The Fickle Teacher)"
What is the main argument in this chapter?
How does it relate to the main arguments in "Thinking About Thinking," "The Wisdom (and Madness) of Crowds," and "Special People?"
I moved the quiz to Thursday because multiple people were confused about the content. The quiz will include questions from Ch. 4-5.
Lessons from Watts, Heffernan, and Meadows for Knowledge Integration
Don't trust something because it seems obvious or appeals to common sense. Never rely on a single person's experience or opinion, especially your own. Try to catch yourself in motivated reasoning.
Learn as much as you possibly can from the people around you. Leverage local expertise. Look for solutions that are already working "on the ground."
Be aware of how you receive information, what kind of information you (do not) receive, and what you could do to improve the information flow. Always be thinking about sampling bias. Seek out non-confirming and negative views.
Adopt a data-driven and measurement-oriented mindset. Be willing to change when presented with good evidence. Update your beliefs and behaviors in the face of evidence. Measure and react.
Develop planning methods that respect the limits of prediction. Scenario analysis and optimizing for strategic flexibility, but more importantly react to the present.
What connections do you see between Watts and the first part of Heffernan?
How does Heffernan define "social capital?"
nodes, edges, whole network, ego network, directed, undirected, isolates, dyads, triads, triadic closure, clustering and closure, tie quality (e.g. type, strength), tie quantity (weight), centrality (e.g. degree, betweenness, closeness, eigenvector), brokerage... and many more.
Looking a little less clustered this year...
Social capital is not an attribute of an individual (like human capital). It is embedded in a network. It's a resource that you have access to. It's who you know, not what you know.
What are some advantages of being embedded in tightly knit groups?
What are some advantages of bridging groups?
Opinion and behavior are more homogeneous within groups than across them, so people who are connected across groups are more familiar with alternative ways of thinking and behaving. Brokerage across structural holes provides a vision of opportunities otherwise unseen.
how many of you are "good writers?"
3 Minutes: How do you normally write?
3 Minutes: Why do you write this way?
3 Minutes: How did you learn to write?
3 Minutes: What could you do to make your writing better?
Now trade with the other people at your table.
"Writing productively is about actions that you aren't doing but could easily do: making a schedule, setting clear goals, keeping track of your work, rewarding yourself, and building good habits." - Silvia
What does Lamott mean by “bird by bird?”
Why are large projects immobilizing?
Why are small assignments useful?
Why do we need systems thinking?
What is "systems thinking" anyway?
A systems ability to survive and persist in a variable environment. E.g. though redundant feedback loops.
Feedback loops that can restore or rebuild feedback loops. learn, create, design, evolve.
the human body, populations, ecosystems
static stability is not resilience.
resilience can be very hard to see unless you exceed its limits
the capacity to make the system structure more complex: lean, diversify, evolve.
simple organizing rules can lead to complex self-organizing systems.
Systems are nested. Higher level systems coordinate the subsystems. this increases resilience and reduces the amount of information that any part of the system has to keep track of. When systems break down, they often split along subsystem boundaries.
Everything we think we know about the world is a model, some of which are better than others. Systems models are especially good models, but they are still imperfect.
"Essentially, all models are wrong, but some are useful."Statistician George Box
Your papers should draw on material from Watts, Heffernan, and Meadows, but not necessarily equally.
I am developing a grading rubric with Bronwyn and Samantha.
Start thinking of as many potentially interesting paper ideas as possible. Record them and have them available for next class.
The rest of the content from Thinking in Systems was covered in in-class activities.
What is a "literature review?"Why are you conducting / writing one?
Just some of the many types of literature reviews: Duke Medical Library and Archives.
Reviewing the literature -- towards any end -- is especially complicated when relevant work is conducted in different disciplinary and interdisciplinary contexts.
Let's assume that your approach is going to be broadly "systematic."
Web of Science, PubMed, JSTOR, Scopus, Google Scholar, Library Databases, etc.
Learn how to read books and articles... (yes...)
Unzip that folder of articles...You have 20 minutes to learn as much about them as you can.
Now, I want you to create piles / clusters of similar articles. You can cluster them however you want. You can create as many clusters as you think is appropriate.
How did you cluster them? Why?What did you learn?
There is another way!metadata
$x$ of science, science and technology studies (STS),scientometrics / informetrics / bibliometrics, etc.
Lots of intellectual diversity!
disciplinary background? non-researcher audience? methods? goal? "computational?" ...
From the first part of class.See slide 5.15
From the second part of class. Recall: brokerage, knowledge, ideas, behavior, etc.
UC San Diego "Map of Science" Data
one-mode social networks (e.g. co-authorship)one-mode content networks (e.g. keyword co-occurrence)two-mode social and / or content networks
Many different goals...
co-authorshipco-investigatormentorshipcommunication...
What do you think a co-authorship network with data from these two papers would look like?
Usually the data should be stored as an edge list and a node attribute file, but there are other ways too.
Giant component from Canadian social science co-authorship network, 2005-2015.
citation and co-citationdocument level, author level, journal level
What do you think citation and co-citation networks would be constructed at the document, author, and journal levels?
Co-citation network of 2015 articles published in Science and Nature.
The nodes in the network have two different classes. Links are between but not within classes.
e.g. article-keyword networks
article-keyword network for articles in Scientometrics and Social Studies of Science from 2005-2015.
Next class: historical bibliometrics and installing software!
To come.