Abstractions: Enabling Data Teams Through Reduced Complexity

Layers of abstraction can drastically reduce complexity, and enable data practitioners to operate efficiently within their data platform, however they can also come at a cost if not done correctly. This talk will explore the methods that can be employed to help promote cross-team collaboration, and reduce the barrier to entry for contributing to the often siloed data platforms by defining a well established framework.

We'll dig into some common practices around building abstraction in the data domain:

- Data modeling
- Building DSLs powered by YAML and JSON
- API layers

Along with some of the common pitfalls:

- Abstracting too early
- Leaky abstractions, exposing technical complexities
- Over abstractions leading to unnecessary complexity