Phil 4.13.17

7:00 – 8:00 Research

  • Reading the HCIC Boaster Poster description. Downloaded to HCIC 2017 folder
    • A “boaster-poster” is a poster that describes your most current research endeavor and/or interest. The idea is to foster dialogue about your topic of interest/research so you can meet like-minded HCIC 2017 attendees. Format for a “boaster-poster” is as follows: a short description of your perspective and interest in this area, plus a description of your work in form of a single page (8.3 × 11.7 inches) poster. Boaster-posters offer an opportunity to showcase the work of new and experienced authors alike. You can use images and text to frame and illustrate your ideas. A list with boaster-poster titles, authors & abstracts will be distributed at the conference, and the posters will be available for view at the HCIC conference. We strongly encourage all student attendees to submit a boaster to HCIC, as boaster authors will have opportunities across the conference to discuss their work with other attendees through a new interactive format for 2017.
      Boaster-poster deadline: June 2nd, 2017
      A pdf that includes:

      • A cover page with
        • Title, author(s) (indicate those available to chat at meeting)
        • At least three keywords
        • A 150 word abstract
      • A draft of your poster
    • So something like “sociophysics-informed design“? I’m thinking that if I can take agent cluster membership and use that to construct a social network graph, I could show something that looks like this: twitterdata1-01
    • Maybe use graph-tool Python library? polblogs_pr
    • Need to look at Zappos and McMaster websites as examples of explorational interfaces
    • Facebook’s guide to handling Fake News. High effort. I wonder what kind of feedback mechanisms there are?

8:30 – 6:00 BRC

  • Sprint planning
  • Doctor visit, 10:15 – 11:00
  • Discussion with Aaron about visualizing high-dimensional clusters in low-dimensional space for intuitive understanding
  • Working through Thoughtful Machine Learning. Very disappointed. The code in GitHub doesn’t match the book, doesn’t even have an entry point, and blows up in the init. Sad! Here’s the offending line (df is the read-in DataFrame):
    df = (df - df.mean()) / (df.max() - df.min())
  • Learning more about the pandas DataFrame here so maybe I can fix the above.
  • Actually, Skillport has useful stuff, but all the videos crash before the end
  • The problem is that the floating point values in the file are being read in as string values, and crashing the calculation. I’ve tried doing an apply function that changes the value but it doesn’t result in the type change. Going to try changing everything to float tomorrow.
  • Helped Aaron break down the tasking for this sprint’s efforts.