Skip to content

Latest commit

 

History

History
42 lines (30 loc) · 2.84 KB

to_do_list.md

File metadata and controls

42 lines (30 loc) · 2.84 KB
Raw

Ways to improve this book/course

  • need to provide more structure and instruction at the outset
  • ultimately, move away from datacamp.
  • Teach python in a more "direct" way for this class. DataCamp teaches too many ways to do something, and too few opportunities to do it. Teach one way of doing the things that need to be done to complete assignments.
  • most specifically, clearly they need more instruction and guidance on thinking about data, drawing conclusions, and showing data in a way that facilitates drawing conclusions.

Videos

Students don't just want the textbook, they want video walkthroughs of the textbook

Tricky Topics

Apparently Jen & Leanne have created a phenomenon. This was asked for by students this year

Visualization

  • definitely need to write a chapter on data visualization. Don't get lost in the weeds of the grammar of graphics before you start, Aaron!!!

  • Matplotlib - don't start with 2 ways of doing things (procedural and OO). Introduce OO and then have a "bonus" section explaining the differences, once OO is deeply ingrained.

  • subplots - explain in more detail, and move description of this from being buried in intro_spike_trains.

    • update all other chapters/sections to use OO approach to Matplotlib

  • in Single Unit -> Working with pandas

    • break into subsections
    • show how to do heat maps

Research designs

  • Separately, a chapter on research designs, such as factorials, and differences between discrete and continuous data.

  • How to critically evaluate EDA. They're not doing this and we're not teaching them how. It's more than just drawing the plots, they need to think about what they're plotting, how to structure the plots to reflect the experimental design and answer the questions being asked.

    • also need to talk about magnitude of difference, statistical thinking a la Geoff Cummings. Including what error bars are and how to draw them correctly, and how to eyeball effect sizes.

  • a clear style guide (chapter) that talks about how to report data - how to write a Results section

Syllabus/Evaluation

  • be explicit that low peer evals will negatively impact project grades. Specifically, team member(s) whose average evaluation from their peers is more than one point (out of 5) lower than the average of the others will have their project grade reduced proportional to their peers' evaluation. E.g., if your team's project received a grade of 80%, but your average score from your peers was 1.3 (out of 5) whereas others' were 4 or greater, your project grade will be reduced as (1.3/5) * 80% = 20.8%.

  • fix rubrics

Demos

  • make individual demos themed - more constraints/less options
  • e.g., first one a review of a YouTube video or blog post
    • second could be a specific function
  • Create a Guide/tips 'n tricks for demos. E.g., USE A MEANINGFUL TITLE, use headers in written docs, etc.