ETL / ELT
ETL: extract data from OLTP sources, transform it into analysis-friendly shape, load into the warehouse. ELT swaps the last two load raw first, transform inside. Generalizes into data pipelines.
What is it?
Extract (periodic dumps or continuous streams), transform (reshape, clean), load. ELT loads raw data first and transforms in-warehouse. Some sources are vendor SaaS APIs (CRM, email, payments), which connector services (Fivetran, Singer, Airbyte) pull in. Reverse ETL pushes analytical outputs back into operational system — the classic case: an ML model trained in analytics and deployed to serve live users (TFX, Kubeflow, MLflow)
What I learned
The pipeline direction matters: analytics isn’t a one-way street anymore — reverse ETL sends trained models back to production, closing the loop.
Learned from
No source linked yet.
Applied in
No practical applications connected yet.