Hi everyone - Join us on Wednesday, Jan 31 @ 10 a.m. PST for a live "Analytics as Code (AaC)" webinar. AaC is quickly changing the data industry as it leverages the power of software engineering principles to drive agility, efficiency, and scalability in your analytics processes: ‎️‍🔥 Automation: Automating solutions via code simplifies deployment and increases efficiency ‎️‍🔥 Collaboration: Collaborative coding platforms like GitHub or GitLab enable seamless teamwork ‎️‍🔥 Governance: With AaC, every iteration of a solution is securely stored and revertible.

Great demo by Andy and Thomas covering critical use-cases. I had a follow-up question, not sure if this is the right channel. Happy to route it to the proper forum. Question: Previously, data analysts on my team would be creating metrics/MAQL and build visuals. However, they would not be so close to VS Code/yaml/code/git process. In that case, if I start maintaining/managing “analytics as code” for better version control, it will always create a hard-dependency for the data engineers to create the metrics, post to server and then become available for the analyst to use. What happens when the analysts continue creating metrics/visuals which are not maintained in the code ? Is there a divergence or conflict during the next deployment ? What is the recommended best practice for this use-case from GoodData ?

Hello Shankar, thank you for the question. Generally, the best practice is to treat your code as the main source of truth. That’s what defining your project. Individual contributors can work visually if they choose, but they should always do so on their dev environment. Once they are happy with the result that they built with Web UI, they “clone” the metadata from server to code (our VS Code Extension can do that) and push the changes to Git. We are considering an implementation when our Web UI can push changes directly to git, but this approach comes with it’s own issues and for the time being we are not going to support it. The only way a new version can get to production is through code and CI/CD pipelines.

Hi Andrii, thank you for the recommendation and good to know that insights created on UI can also make it to the code base. I will explore on this and reach out in case of any questions. Is there a channel where I can route my queries concerning this feature ?

We do not have a dedicated channel for GoodData for VS Code as of yet, so feel free to write to #gooddata-cloud