ReactPythonFastAPIApache AirflowSnowflakeKubernetes

DataWeave

DataWeave lets non-technical analysts build and schedule data pipelines without writing code. It connects to REST APIs, SQL databases, Google Sheets, and cloud data warehouses, providing a live preview of transformations before execution.

680 GitHub stars
DataWeave - main screenshot

// key outcome

Enabled analysts to own their own data pipelines, reducing data team tickets by 60%.

01 / Problem

The Problem & User Context

Data analysts needed custom ETL pipelines but had to file tickets with the data engineering team for every new source or transformation, creating a 2-week backlog that blocked business decisions.

02 / Solution

Approach & Tradeoffs

A React Flow-based canvas lets analysts connect nodes representing sources, transforms, and destinations. A Python/FastAPI backend validates and executes pipelines via Airflow. Real-time cell previews at each node eliminate guesswork.

03 / Architecture

Technical Architecture

React + React Flow for the canvas. FastAPI + Pydantic for the execution API. Apache Airflow for scheduling. Snowflake as the primary data warehouse target. Kubernetes for horizontal scaling of worker pods.

// tech stack

ReactPythonFastAPIApache AirflowSnowflakeKubernetes

// architecture diagram

Add your architecture diagram here (Excalidraw, Figma, or SVG)

04 / Visuals

Screenshots & UI States

DataWeave screenshot 1
DataWeave screenshot 2
05 / Results

Results & Lessons Learned

  • 60% reduction in data engineering support tickets
  • Average pipeline creation time: 15 minutes (vs. 2-week queue)
  • 40+ connectors including Salesforce, HubSpot, Stripe, and BigQuery
  • Processing 500M+ rows per day in production

Explore the Code

All source code is open-source on GitHub. Issues, PRs, and feedback are always welcome.