gvdr
commented
6 years ago Sorry for the late answer. And THANKS for the feedback!
The "Carpentry" stuff comes from Ian Bogost's idea that "philosophy" is building stuff. I would like my students to discover a bit of "Object Oriented" (ontology/feminism). The core take home message is that everything, also models, algorithms, data, are "objects", and we never fully understand them. We can explore and investigate them by inter-acting with them, but our knowledge is always limited.
There are a couple of practical consequences: 1) do not believe that your model fully capture something or somebody (i.e., do not reduce a person to their representation in your model) 2) plot your data and play with your models in exotic ways if you want to understand more what they are doing: think about the Anscombe's quartet or The Datasaurus Dozen, you can interpret those either as "strange data" or as a limitation of the diagnostics we use for linear models.
wrightaprilm
Is the idea of this segment to discuss documenting the workflow, for how the data were used? I don't know some of the terms that are on the syllabus, but that's what I'm picking up as a core component - how can a stakeholder verify that what has been done with the data is what was agreed to? With facebook, there's so much obfuscating language, third-party agreements that this is nearly impossible, but for smaller scale applications, that may not be the case.