| layout |
|---|
lesson |
This module will cover creating environments for reproducible science
Audience may have little or no experience with programming, but may be proficient at button clicking. (mostly coming from psychology/neuroscience phd programs)
- expected size of audience: (10-20)
- time investment (25 hours?) + 4 hours in the brainhack
Battlestar Galatica is a second year graduate student measuring the weight from two groups of rats on a weekly basis. One group of rats has aperol spritz with every meal, and the other has water. After each week, she types the data into a excel sheet, gets descriptive statistics of the samples for that week/overall and runs a t-test at each of the time points. she already has an excel sheet that auto-populates the fields, but she is thinking about doing more advanced analyses of the data (longitudinal mixed effects models). She has not worked with python/R before, and would like to start out small to replicate her excel workflow in R/python, with a longer term eye on implementing the more advanced model.
- learn computational vocabulary
- use fundemental tools
- interact with people with different backgrounds
I will be hosting a weekly flipped classroom for 5 weeks prior to the workshop. I will ask the participants to begin to think about what repetitive tasks or analysis steps they perform in excel and write down (and share) their goal for using shell/python/R to (partially) automate the task. I will assign the lesson for the participants to go through on Monday, and host in-person office hours on Friday (3:00-5:00pm). At the office hours, participants can ask about roadblocks they had in the lessons, or in their automation goal. Throughout the week, those that have registered will have access to a mattermost(?) channel/or direct messages to ask questions to me/each other as they engage with the lessons.
| Week | Topic | objectives | module |
|---|---|---|---|
| 1 | shell/installfest | (move/rename/cp/rm files) (install github-desktop/appropriate bash environment/conda/R) | http://swcarpentry.github.io/shell-novice/ |
| 2 | git/github | (understand conceptually what git accomplishes) (syntax for adding/commiting/pushing/forking) | http://swcarpentry.github.io/git-novice/ |
| 3 | python | (interacting with tabular data numpy/pandas) (create conda environment) (some data structures/for loops) | http://swcarpentry.github.io/python-novice-inflammation/ |
| 4 | R | (interacting with tabular data) (some data structures/for loops) | http://swcarpentry.github.io/r-novice-inflammation/ |
| 5 | hpc | (how to submit jobs to cluster) (use a singularity container) | (make some material) https://hpc-carpentry.github.io/hpc-shell/ |
After week five, we will be hosting a 2-day brain hack with two 2-hour tutorials. While the participants will have say in what exactly is covered, my plan if no preference is indicated:
-
creating environments (dockerfile/singularity)
-
testing your analysis (travisci/circleci)
- They create a script that solves their problem (potentially as a personal project at the brain hack)
- the script is version controlled
- the script is on github
- the script has a test
- the script runs in a container
- the script is automatically tested
- They interact with people outside their field (by attending the brain hack and acting as a source of a user profile for a tool developer)
- they become interested in reproducible (efficient) practices
- do I introduce a text editor/which one do I introduce?
- what debugging strategies should I introduce?