Skip to main content

1. Getting Started

This page covers the essential setup required for both the Data Owner (DO) and the Data Scientist (DS) to join the secure, distributed data science ecosystem.

1.1 Data Owner (DO) Setup: Hosting the Datasite

The DO prepares their machine to securely host the data and the RDS Dashboard.

StepInstructionCommand/Action
Install CLIInstall the SyftBox Command Line Interface for client interaction.curl -fSL https://syftbox.net/install.sh | sh
AuthenticateLog in with your email to receive an OTP and obtain your refresh token.cat ~/.syftbox/config.json
Deploy DashboardRun the RDS-Dashboard Docker container, using your email and token as environment variables.See DO Quickstart for the full Docker command.
Access DashboardVerify the dashboard is running.Navigate to localhost:8000
Upload DataCreate a dataset container by uploading your private data and accompanying mock data.Dashboard > Datasets > Add a Dataset

1.2 Data Scientist (DS) Setup: Preparing the Client

The DS prepares their local development environment for writing and submitting Federated Learning jobs.

StepInstructionCommand/Action
Install CLIInstall the SyftBox CLI locally.curl -fSL https://syftbox.net/install.sh | sh
Install DependenciesInstall the required Python packages (e.g., Flower, Pytorch, Syft Flwr).pip install syft_flwr (plus project-specific dependencies)
Clone ProjectGet the source code for the FL job you intend to run.git clone [project_url]
Client InitializationInitialize secure connections to your client and the DO's datasite.See DS Quickstart for Python code block.
Validate LogicUse the mock data to test your model's code structure and logic without accessing private data.Run FL workflow locally on mock data.