Mastering Essential Skills: Incorporating Software Development Principles in Statistical Practice
New Zealand Research Software Engineering Conference 2023
18 September 2023
What I do as a statistician
Software engineering skills I wish I learned about
What I learned at uni:
statistics
data analysis
programming in R and Python
What I wish I had learned:
Version control
Managing your computational environment
Documentation
Unit testing
Workflow management
Why do we need software engineering skills?
Sustainability
Open science
Version control
Used under a CC-BY 4.0 licence. DOI: 10.5281/zenodo.3332807
Code versioning with Git
Access to the history of the project
Easy collaboration on code
Standardised way of sharing and publishing code
Data versioning
- Data are not static
- Link results to versions of the data
- Versioning raw and processed data
Managing computational environment
Computational environment matters
Dependency management tools
Use dependency management tools like renv (R) or conda (Python) in your work
1. Record dependencies
2. Isolate project library
Documentation
Write it down!
Codebooks/data dictionaries:
Literate programming
Concept: mix code and plain text
Record reasoning, notes and interpretation alongside source code and results
Reduce copy-pasting for reports and presentations
(Unit) testing
Testing to catch issues
Write formal tests for:
data (dimensions, range of values, etc.)
code (correct output type, handles errors)
analysis steps (sensible results, returns known answer)
Using simulations to test statistical methods
Test models and algorithms with simulated data
Is the model appropriate to answer the research question?
How do assumptions violations affect the results?
Workflow management
A real-life example of a complex analysis
Multiple scripts, folders, datasets
Interdependent analyses
What to do when a part changes
Workflow management tools
Turn your analysis into a pipeline, i.e. series of steps linked through input/output, which will be executed in the correct order
Conclusions
Statisticians and data scientists need more than statistical skills
Software engineering practices are crucial for work sustainability and for collaboration
These skills should be taught in statistical degrees (but there are lots of resources out there!)
To go further
Richard McElreath’s talk about Science as Amateur Software Development
The Turing Way handbook to reproducible, ethical and collaborative data science
Bruno Rodrigues’ book on Building reproducible analytical pipelines with R
The Good Research Code Handbook by Patrick Mineault
Presentation disclaimer
References
