Trust & Security Overview
.png)
DataUniversa enforces dataset-level access control, structured dataset validation, full audit logging, and controlled file handling across all system interactions.
All datasets pass through an admissibility and validation process before access is permitted, while every critical action, including submissions, downloads, reviews, and denied access attempts, is logged with contextual metadata for traceability and investigation support.
Access is role-controlled and restricted at the dataset level, files are privately stored and delivered through backend-controlled routes, and all user actions are tied to authenticated identities.
The result is a system designed to provide controlled dataset exposure, auditable activity tracking, strong separation between users and data, verified accountability, and measurable operational oversight rather than assumed trust.
1. Access Control
- Role-based access (Admin, Reviewer, User, Buyer)
- Dataset-level permissions enforced via centralized service
- Buyers can only access approved datasets
- All access decisions are:
- evaluated before execution
- logged (granted + denied)
2. Dataset Validation
- All datasets pass through structured audit:
- provenance validation
- structural checks
- admissibility criteria
- 5-month audit completed on primary datasets
- Dataset lifecycle controls exposure:
- pending → restricted
- approved → accessible
3. Logging & Auditability
Every critical action is logged:
- dataset submission
- dataset viewing
- file download
- access denied
- audit/review actions
Logs include:
- user_id
- dataset reference
- IP address
- action + severity
- metadata
4. Data Isolation
- Users only access their own datasets
- Buyers access only approved datasets
- Admin/reviewer access controlled via roles
- Files stored privately (no public exposure)
5. File Security
- Files stored in AWS S3 (private)
- Access only via backend-controlled routes
- No direct public file links
- All downloads logged
6. Authentication
- Google OAuth (Socialite)
- Email + password login
- All dataset submissions tied to authenticated user
7. Incident Handling
- Access anomalies logged and classified
- Unauthorized access attempts tracked
- Investigation supported by full audit logs
8. AI Handling
- AI module currently restricted
- No sensitive dataset data sent externally
- All AI interactions logged
9. SOC 2 Status
DataUniversa has implemented SOC 2—aligned foundational controls across access, logging, and data handling. Formal SOC 2 audit preparation is in progress.
We enforce dataset-level access control, log all access attempts, and validate datasets through an internal audit layer. Here's what that looks like in the system.
Administrative visibility into dataset access, downloads, and security events.
All unauthorized access attempts are blocked and logged with full context.
All dataset access and file downloads are tracked and auditable.
Dataset submission integrity and traceability.
