Andreas Weigend, Social Data Revolution | MS&E 237, Stanford University, Spring 2011 | Course Wiki

Class_06: Catch Up

Date: April 14, 2011
Audio: weigend_stanford2011.06_2011.04.14.mp3
Other: (#I cannot seem to open the two links above)
Initial authors:[Asha Gupta,], [Tim Holley,], [Becky Nixon,]

Key Points

  • Reviewed what we learned in DF1, HW1
  • Overview of HW2 on Influence
  • Discussion of the Social Data Lab

Dog Food 1

  • Jason will be posting some of the top suggestions on the Wiki
  • Examples of submitted ideas:
    • How many conversations do you have face-to-face with people in a day?
    • Poll people at the beginning of the day and at the end of the day, ask how many distinct smiling faces they’ve seen lately and describe demographics
    • Ask people to rate their happiness at different times of the day, examine what factors affect happiness: when they’re hungry, what country they live in, etc.
  • Discussed differences between unaided feedback and aided feedback when surveying - E.g., When surveying people asking open-ended questions often yields different results compared to asked multiple-choice answers
  • Recommended reading: “Telling more than we can know
    • E.g., Dental Floss Survey
    • Divide people randomly into two groups
    • Ask how often they floss their teeth
    • Group A options per day
    • Group B options per week
    • Result: Those in Group A are more likely to sign-up for dental hygiene

Dog Food 2

Doubling Effect and Pace of Technological Change

  • People work to invent things to overcome today’s problems
  • Pace of technological change is accelerating
external image 7lzfz23s7DK_Jcm4XW4IXsFsOuglcQo8h5__AgNuw4hscDNmHGhRecehLhZt7NnSROjlDlZ8TvJNrc5keeDduw54gXYTZzgbE3t54X0wXjsAhfEx3A

Information Overload

external image 4qEKGkatEPovylb_naRZBzu7XtpwSqehkcmOGEI3wXwcCd60fz-pN6q6L-o8gFBnzN7ZjlXaiLsZFgrd_m7tjtopNDbr14-t7P9-5KhO03H1d5jsJQ
    • There are so many forms of information competing for our attention - emails, message popups, newsfeeds, RSS feeds etc.
    • Today, distribution costs are negligible (such as retweeting), what is scarce is our time; there is a large cost to interruptions now
    • Further, more data is not necessarily benefiting us:
    • It would be better if we could associate metadata with retweets - such as an integer to signify how important someone considers a piece of information
    • We need to come up with ways of organizing information - perhaps by time line, grouping similar pieces of info together.
    • An interesting quote regarding information overload : "Data is like food. A good meal is served in reasonably-sized portions from several food groups. It leaves you satisfied but not stuffed. Likewise with information, we're best served when we can partake of reasonable, useful portions, exercising discretion in what data we digest and how often we seek it out."
    • William Van Winkle - William Van Winkle, who is a freelance writer, who specializes in technology's social impacts, argues that technological innovations (such as e-mail, voice mail, phone calls, meetings, business journals, faxes, memos, manuals, Web research) have overall increased, not reduced our work load, and have reduced our productivity. For more on this topic check out his article here ; he also gives some suggestions on how to gain control of the information in our lives.
    • A very cool video that lists interesting facts about the world around us, the effects of data overload in the past 10 years, and how technology is changing our lives, and our futures: YOUTUBE VIDEO
    • Nicholas Carr suggests that the internet is causing a fundamental shift in the wire that our brains wire themselves due to the constant context switching: Wired Story

Psychology of Clicking

external image L6XfuDcP3KK0TFfEz_kwZORy998WHQgMNKQhOoqmo7K8C7-8hTY1AyEVQeHreG78fAhgbZwDwD00PEGsUZ4VAVanjr2Nfrz-uVNr_l89LD5W3tOBBw

Translating Language on the Web

  • DuoLingo: Learn a language while translating web pages
  • Interesting insight: In early 90’s IBM changed language translation from rule-based to statistical-based, if we have enough data we can use analysis without actually understanding what the language means.
  • Chinese room theory: If you can carry on an intelligent conversation using pieces of paper slid under a door, does this imply that someone or something inside the room understands what you are saying?external image EJ0QErqF2OQ2EZ0o5gDTh2uPNDlJ9CSl3D2IsFAnw6oP68QkEbyz0nV2fg6H-VWbl8N0ouW2AQ9yP6A0KlTvBY68-2zmmRvsH8dm0GVgogP7Lkr9HQ

Dog Food 3

  • Assigned today, due Tuesday, April 21 at noon
  • Make an account on Quora, look around and understand the culture
  • Here are some articles on Quora (found using Quora) :)

external image 8aY6sta5auMnx_1tsvyjBUUgdfbYrwPYgosAVhcJGUfELAEp2mubTBtpYxuQrMTTtgaqiwBZWjBlxEmdp1E2EciOua8JaPwEPFViCmpR5tR90L9dSw
  • Follow people and topics, get followed by others, answer questions, vote answers up and down
  • Ask a question you are genuinely interested in and see what happens
  • Don’t write long stories in your response, just give a few key insights

HW 2

  • Read chapter 4 of Mining The Social Web
  • Read content from Duncan Watts & Sinan Aral to get more information on influence, build your own mental model
  • There are a few questions to keep in mind as you read the articles:
    • What really is influence? Someone that grows the network or helps others to engage more in the network? Influence means getting others to perform an action?
    • What could potential metrics be?
      • Time spent reading web pages is often used (erroneously) as a measure of engagement.
      • People run algorithms on mean of time spent on page, but data is often not useful because people leave their browsers open while they aren’t actively engaged
      • At Amazon they measured how far down a user scrolls down the page. This gives you an idea of how much they are actually reading/viewing.
      • Short-click vs. long-click
  • Companies measuring influence

external image MXtp7FLe9Q2Ywb1KdhmajJdB7sJe5IbUIXk2fUnhu7mHDFYCRQKVTCOeOCRP8a_YiH9X83ysCuKg9PH-e4sqB5nc5guelX7HkO_0cSyRN0Sg8fLZeQ
    • SocMetrics -- SocMetrics tracks key influencers with the purpose of eventually connecting businesses to them. They gained a lot of publicity at SXSW by tracking which parties the most influential people were going to.
  • PeerIndex is another company that is trying to create the de facto metric for measuring influence online.

  • The total number of followers a person has does not necessarily translate to influence in social networks like Twitter. Check out this article and the analysis conducted by Twitalyzer for more information.
external image moz-screenshot-3.png

  • Influence is coupled with social capital

Extra Credit

  • There is no project this year, however there are extra credit options to get an additional 20 points
  • If you have a great idea for a project, you can drop one of the homeworks and do a project instead

Homework Grading

  • Jason has graded about half of homework #1 and will return grades but not individual comments
  • Whenever he finds an interesting idea or particularly good example he will put it on the wiki, you can compare your own content to the examples on the wiki
  • If you have individual grading questions ask Jason

Social Data Lab

  • 5 companies interested in social data opportunities will sponsor the lab
  • Former students and current students will participate
  • The lab’s end goal is to have a full day event in May, they will pick 2 topics to discuss and innovate on, at the end of the day they will all go to dinner
  • Participating has no influence on your grade in the class
  • Attitude is the most important ingredient
  • There are no coding prerequisites
  • One sponsor will be Swift – they are extremely interested in understanding identity
  • What is the goal of the lab?
    • The goal for you is to learn about real problems that real companies have
    • The goal for the companies is to get ideas & innovations from you
    • Fair exchange – Andreas is best at bridging


  • Next Wednesday, April 20th a group of Knight Fellows will talk about the future of journalism
  • Monday, April 25th, DFJ is doing an event called influencer series – VC angle on social data
    • RSVP so Andreas knows roughly how many people are coming