Code owners
Assign users and groups as approvers for specific file changes. Learn more.
reproducibility.md 2.31 KiB
Reproducibility
- ensures credibility
- key requirement for follow-up and collaborative studies
Why is our workflow not reproducible?
Lack of provenance:
- Input data downloaded from “some website”
- Copy & paste operations
- Manual text entry
- Analysis not coded
Reproducibility
Learning to code workflows and analyses
Spreadsheets alone
- Is great for looking at data.
- Data entry is fast.
- Analysis flow is hidden and not in focus.
Coding
- Is great for controlling analysis
- Data is hidden.
- Flow is visible.
Develop data science skills
- Develop good data management and analysis habits.
- Start coding your analysis within spreadsheets.
- Make yourself familiar with a statistics environment such as R, Python or Matlab
- No need to learn a high level programming language such as C++ or Java.
Table
"Tabular format of data"
Header
- one line!
- good names of columns
Rows
- represent observations/entities
Columns
- represent property of the observations
- one data type
