Skip to content

Latest commit

 

History

History
104 lines (90 loc) · 4.68 KB

design.md

File metadata and controls

104 lines (90 loc) · 4.68 KB
layout subtitle
lesson
Lesson Design

Stage 0 - Assumptions

  • Audience
    • Graduate students (PhD students, postdocs, technicians) in numerate disciplines from cosmology to economics
    • Who have used the terminal or at least attended shell-novice
    • Who are literate in any given programming language (python, C/C++, Fortran, ...)
  • Constraints
    • One full day 09:00-17:00
      • 06:30 teaching time
      • 1:00 for lunch
      • 0:30 total for two coffee breaks
    • Learners use native installs on their own machines (ssh session)
      • May use logins to local cluster
    • Dependence on other Carpentry modules
  • Exercises will mostly not be "write this code from scratch"
    • Want lots of short exercises that can reliably be finished in allotted time
    • So use MCQs, fill-in-the-blanks, Parsons Problems, "tweak this code", etc.
  • Lesson materials
    • Notes for instructors and self-study will be written in Markdown
      • We've tried writing/maintaining lessons as Notebooks...
    • Learners will be provided with one Notebook per episode containing exercises

Stage 1 - Desired Results

Goals

  1. Get learners to the stage decribed in the "Software" section of "[Good Enough Practices in Scientific Computing][good-enough]".
    • Goals
      1. Make it easy for people (including your future self) to understand and (re)use your code
      2. Modular, comprehensible, reusable, and testable all come together
    • Rules
      1. Every analysis step is represented textually (complete with parameter values)
      2. Every program or script has a brief explanatory comment at the start
      3. Programs of all kinds (including "scripts") are broken into functions
      4. No duplication
      5. Functions and variables have meaningful names
      6. Dependencies and requirements are explicit (e.g., a requirements.txt file)
        • This rule is not covered in this lesson
      7. Commenting/uncommenting are not routinely used to control program behavior
      8. Use a simple example or test data set to run to tell if it's working at all and whether it gives a known correct output for a simple known input
      9. Submit code to a reputable DOI-issuing repository upon submission of paper, just like data
        • This rule is not covered in this lesson
  2. Enable them to make sense of other onlines tutorials and resources

Summative Assessment

  • Midpoint: plot bar chart showing average GDP per continent
  • Final: debug and extend a short multi-function program to handle data laid out differently

Essential Questions

How do I...

  • ...read, analyze, and visualize a tabular data set?
  • ...process multiple data sets?
  • ...tell if my program is working correctly?
  • ...fix it when it's not?
  • ...find and use software other people have written instead of writing my own?

Learners Will Be Able To...

  • Run code interactively
  • Run code saved in a file
  • Write single-condition if statements
  • Convert between basic data types (integer, float, string)
  • Call built-in functions
  • Use help and online documentation
  • Import a library using an alias
  • Call something from an imported library
  • Read tabular data into an array or data frame
  • Do collective operations on arrays and data frames
  • Create simple plots of data in arrays and data frames
  • Interpret common error messages
  • Track down bugs by running small tests of program modules
  • Write non-recursive functions taking a fixed number of named parameters
  • Create literate programs in the Jupyter Notebook

Learners Will Know...

  • That a program is a piece of lab equipment that implements an analysis
    • Needs to be validated/calibrated before/during use
    • Makes analysis reproducible, reviewable, shareable
  • That programs are written for people, not for computers
    • Meaningful variable names
    • Modularity for readability as well as re-use
    • No duplication
    • Document purpose and use
  • That there is no magic: the programs they use are no different in principle from those they build
  • How to assign values to variables
  • What integers, floats, strings, and data frames are
  • How to trace the execution of a for loop
  • How to create and index lists
  • How to trace the execution of if/else statements
  • The difference between defining and calling a function
  • What a call stack is
  • Where to find documentation on standard libraries
  • How to find out what else scientific Python offers