The Three Environments

(Image Source: Miami University)

Dev

• The development environment is the product sandbox
• Most “data science” happens here:
  • Data analysis & modeling
  • App prototyping
  • ETL

Comparing dev for data science vs dev for software engineering:

	Data Science	Software Engineering
Goal	Explore relationships in data that may develop into live products	Build & implement a specific feature for a live product with pre-defined requirements
Tools	“Fully fledged” data science IDE (RStudio, VSCode) can encompass Dev, Test, & Prod	Dev, Test, & Prod are differentiated by environments & containers

Most of what data scientists do doesn’t end up as a live product in the state it was created; these entities think differently! (Source)

Test

• Test is for testing :)

• Tests used for many reasons incl. security, portability, performance, usability

Prod

• Gold standard (where your live product is released into the wild)
• Should be guarded by Continuous Integration/Continuous Deployment (CI/CD)
• Ideally zero manual interaction and zero changes to the actual code