Data Studio User Guide

Create projects, manage datasets, design pipelines, label and curate data, and connect data workflows to Synthex, training, and evaluation.

Who This Guide Is For

• Data engineers
• Data scientists
• Labeling teams
• ML engineers

Where To Go

Page	Use It For
/data-studio	Data Studio dashboard.
/data-studio/new	Create a data project.
/data-studio/datasets	Manage datasets.
/data-studio/pipelines	Build data pipelines.
/data-studio/[projectId]	Open a project workspace.

Core Concepts

Concept	Meaning
Project	A workspace for a data objective, dataset group, pipeline, or labeling initiative.
Dataset	A versioned collection of files, records, labels, and metadata.
Pipeline	A repeatable sequence for ingestion, cleaning, validation, enrichment, or export.
Labeling workflow	A review and annotation process for supervised learning or evaluation data.
Integration	A connected data source, storage backend, or downstream consumer.

Common Workflows

Create a dataset project

1. Open Data Studio and create a project.
2. Import or connect data.
3. Review schema and quality.
4. Create a cleaning or transformation pipeline.
5. Label or curate records if needed.
6. Export to Synthex, Model Training, or Evaluation.

Build a reusable pipeline

1. Open Pipelines.
2. Choose source and destination.
3. Add validation, transformation, and enrichment steps.
4. Run a sample.
5. Fix errors.
6. Schedule or save the pipeline.

Best Practices

• Keep raw, cleaned, labeled, and exported datasets versioned separately.
• Record source ownership and retention requirements.
• Use validation checks before exporting to training or evaluation.
• Document label definitions and reviewer expectations.