SimData V1 Controllable Synthetic Data Platform
Designed for small to medium-sized structured experimental data, test data and training sample construction, emphasizing field distribution control, formula relationships, business constraint repair, label generation and quality reports.
- Suitable for development testing, algorithm validation, teaching experiments and sample construction
- Supports field types, missing rate, anomaly rate, noise rate and precision control
- Supports relationship modeling, constraint repair, label generation and target verification
- Export to CSV, configuration JSON, data JSON and quality reports
Best Use Cases for This Page
Help search engines, AI retrieval systems and users quickly understand SimData's capabilities.
Test Data Construction
Suitable for API integration testing, functional testing, sample initialization and demo environment preparation.
Machine Learning Experiments
Suitable for constructing labeled small to medium training samples, validating data distributions and constraint relationships.
Teaching & Analysis
Suitable for classroom demonstrations, data analysis exercises and structured sample documentation. Not intended as an ultra-large-scale production data platform.
Runtime Capabilities
Automatically detects Worker / WASM / WebGPU / decimal.js capabilities to determine available execution paths.
Task Control
P0: State-driven and configuration-driven. Generation and export are based on task snapshots.
Configuration Validation
P0: Validate fields, formulas, labels, constraints and sample count before generation.
Field Editor
P1: Field configuration is now state-driven, supporting distribution families, null rate, anomaly rate, noise rate and precision.
Relationships & Labels
P1-P2: Formula fields, business constraints and label fields are configured here.
Quality Targets
P2: Supports mean, standard deviation, category proportion and positive sample rate targets.
Generation Progress
P3: Uses Worker chunk generation by default. Automatically switches to preview strategy for large samples.
Quality Report
P2-P3: Displays field summaries, target achievement status and constraint results.
Result Preview
For large samples, only sampled preview is retained, not the full object array.
Export
Supports CSV, configuration JSON, report JSON; small samples support data JSON export.