Code Review in Mode
Curious if anybody has a way they do code review with Data teams in mode.
We're a small startup team and only I really live in mode spinning up analytics. Has anyone found an effective way to code review SQL for precision and accuracy?
Apologies for the delayed response. All of the process around keeping analytics code in good shape (including code reviews) is hard to do well, and we at Mode certainly empathize with this challenge. Best-in-class organizations that we've seen use a combination of tools and processes in order to solve this challenge:
- A cloud-based, collaborative platform for analytics exploration and delivery (such as Mode).
- Open source tools like DBT ( https://www.getdbt.com/ ) to turn our core analytics code into something extensible, transparent, and durable. Our production analytics are all written/managed in DBT and we host our DBT code in GitHub repos that are subject to the same kind of eng-style workflow and checks and balances.
- Internally, we have a number of folks who act as peer-reviewers for any analytics code before it goes into our production DBT environment. We regularly look at the tools and reports that our business uses to decide what bespoke code gets absorbed and promoted into the DBT "production" stack.
Here are a few resources that might be helpful when thinking about how to design this workflow: