What’s considered best practices is constantly evolving, as new challenges emerge, and software change and automate the workflows of data teams.
When Looker introduced a modelling layer with version control, it was a game changer for many data teams. When DBT introduced testable model generation inside data warehouses, it changed things againt.
In 2024, data teams have more powerful tools at their disposal than ever before. These are the best practices I think every data team should implement:
- Version control
- Testable models
- Quality control metrics
- Data dictionaries and metadata
- Granular access controls and permissions
In addition to these ones, I also predict the following shift for 2025: the full stack data analyst. With all the powerful tooling for ETL jobs, data warehouses and dashboarding, there’s no longer any reason to separate data engineering from data analytics.
Better tooling and AI means self-serve will become even more common, even for more complex tasks. Data analysts will do fewer dashboards, and focus more on a strategic research, organisational training and data engineering.
Leave a Reply