Class 3

Andreas Weigend | Social Data Revolution | Fall 2016
School of Information | University of California at Berkeley | INFO 290A

Video: Social Data Revolution: Topic 5
Transcript: sdr2016topic05.docx
Audio: sdr2016topic05.mp3

WikiLead: Brendan Mackie

Contributor: Zahra Hadi

Topic 5: Context and Conditions

Guest Speaker: AnnaLee Saxenian, Filippo Rossi


Topic 5: Today we're looking at how social data can 'read minds.' Context and Conditions focused on how we can use technology to learn about the human emotional state, and what we can do with the data we collect.

Micro expressions and the reading of emotions
Apps can look at micro expressions to see what people are 'really' thinking. Does this by looking at people's faces and attempting to read the seven basic emotions. One of the problems is that we don't experience just a single emotion at once, but usually feel some combination of emotions.
Amusement, Contempt, Contentment, Embarrassment, Excitement, Guilt, Pride in achievement, Relief, Satisfaction, Sensory pleasure, and Shame. (Not sure which seven they use.)

Can you cheat these methods? No, because they don't read macro expressions, instead they look at micro expressions--miniscule movements of the facial muscles that are difficult to identify by the naked eye and hard to consciously control.

Who buys this stuff? Big companies have used to look at employee satisfaction while using particular pieces of software.

What would we do if we had access to this real-time data on people's emotions
  • You can get people's reactions to big events, e.g. Apple announcements of big products
  • A/B testing of ads, products and so on, using people's emotional reactions rather than formal surveys
  • "Big data-fy" and gamify emotions

A backwards glance over the class with the Dean
What have we learned, what is surprising?
  • Questions about privacy
  • People are surprisingly okay about corporations collecting our data. "I don't trust Facebook anymore."
  • There are protections for how people use data
  • We need new frameworks for what privacy is now--what are they?
  • There are millions of ways to crunch our data

Over the past decade, has the balance of power shifted more towards people, or more towards organizations?
People who say corporations:
  • Because they have so much data
  • We don't know how important our data is and we don't need to use it

People who say individuals:
  • We are creating the content; they have control of it--but we are the content
  • We don't have the profit incentive to use our data

How do we learn?
  • The book is interesting, but it sinks in when we talk about it
  • Because it's a diverse student body, we have a wide range of opinions leading to good discussions
  • As this is uncharted territory, we bring up more questions than we answer

Odds and Ends
The Marshmallow Test:

A legendary psychological experiment testing self control. Give a kid a marshmallow. Tell her that in ten minutes, if she does not eat the marshmallow, she will get another marshmallow. Go into a room with a two way mirror. Wait and watch. It's great! The finding is that self-control is fairly robust--we have the same self control at age ten, twenty, thirty, and so on. Surprising outcome is that people with greater self control have better life outcomes--higher incomes, longer lifespans--than people with less self control.


Seven Basic Emotions

Andreas Weigend | Social Data Revolution | Fall 2016
School of Information | University of California at Berkeley | INFO 290A