This is a companion discussion topic for the original entry at https://blog.rstudio.com/2021/09/16/the-advantages-of-code-first-data-science
Photo by Christopher Gower on Unsplash
RStudio has worked with hundreds of different data science teams, and we’ve seen three key strategies that help maximize their productivity and impact:
- Adopting open source as the core of their work
- Leading with a Code-First approach
- Implementing a centralized data science infrastructure
Collectively, we call this approach Serious Data Science. In this post, we focus on the benefits of a Code-First approach.
A no-code approach to data science has some serious drawbacks, as described in this video:
As we discussed in depth in a recent webinar, a Code-First approach is important because:
- Code provides the flexibility to build and share the most valuable insights, tailored to the analytic problems and needs of your stakeholders
- Code enables fast iteration and updates
- Code by its nature is reusable, extensible, and inspectable
Code-First helps overcome the pitfalls of no-code approaches, as shown in the table below:
| No-Code Problem | Code-First Solution |
|---|---|
Difficulty in tracking changes and auditing work |
Code, coupled with version control systems like git, can track what changed, when, by whom, and why. Code can be logged when run for auditing and monitoring. |
|
No single source of truth |
Centralized tools can create a single source of truth for data, dashboards, and models. Version control can track multiple versions of code separately without creating conflicts. |
Difficulty in reproducing and extending work |
Code can enable reproducibility by explicitly recording every step taken. Open-source code can be deployed on many platforms and is not dependent on proprietary tools. Code can be copied, pasted, and modified to address emergent problems as circumstances change. |
Limitations on analysis techniques and presentation formats |
Code can allow you to analyze and present all your data as you need to in the form of custom dashboards and reports. Code can pull in new methods and open-source work without waiting for vendors to add proprietary features. |
To learn more
If you’d like to learn more about a code-first approach to data science, you can watch our recent webinar here or read an overview of the webinar in this blog post. For a broader view of Serious Data Science and links to more resources, see this page.