cwnote_2014_04_19T08_56_26

background
outline

data science

  practices (managerial)

    reframing questions as ML
    better wrong than "nice"
      better science:
        examples: not even wrong
        hairballs
        BCS ( http://en.wikipedia.org/wiki/Bowl_Championship_Series )
      better career
    be relevant
      vs. look where light is good
      focus on what you value, rather than p-values
    aim for hypothesis vs data jeapordy
      (however, sometimes you fight with data you have not the data you want)
    befriend experimentalists
      donee trouvee
      even einstein needed them to change the narrative
      happens in biology too

  skills and workflow (functional literacy)
    find quantifyables  
    strawman first
      sometimes ML leads to EDA
    small wins before feature engineering
    data engineering before data science

  culture
    be communicative (rhetorical literacy)
      [also a skill: strive for predictive yet interpretable]
    be skeptical (critical literacy)
      esp. of yourself
    be empowering
      ebfret/wigginslab code more generally
    be transparent
      more generally...
    promote literacy
      (better) promote multiliteracy
        [cf. Selber, Stuart A. Multiliteracies for a Digital Age. Illinois: Southern Illinois University Press, 2004.]
        functional literacy
        critical
        rhetorical

get involved!
  nyt: hiring!
  lean workbench
    develop!
    alpha test!
    sutents+postdocs!
  columbia: hiring!
    postdoc in EHR+ML!