Data Collections
Package your plant tags into a curated, shareable dataset — then publish it as a clean API surface for external consumers, without ever exposing your internal tag identifiers.
What Is a Data Collection?
A Data Collection is a named, curated group of measurement tags from one of your plants. You choose which tags to include, assign each one a consumer-safe alias, optionally add descriptions and display ranges, then publish it as an active data feed.
Publishing requires that every tag has an assigned alias — the API cannot expose tags without a consumer-facing name.
External consumers (partners, analytics tools, dashboards) access the collection through the API using the aliases you defined. They never see raw internal tag names such as FI-101 or TI-204.
Plant Tags (FI-101, TI-204…)
→ Data Collection (Curated · Aliased)
→ External Consumers (Partners · APIs · Dashboards)Key Concepts
| Concept | Visibility | Description |
|---|---|---|
| Tag | Internal only | The raw measurement point in your plant system (e.g. FI-101). Never exposed to external consumers. |
| Alias | Shared with externals | Consumer-safe snake_case name for a tag (e.g. total_mill_feed_rate). Must be unique within the collection. |
| Description | Shared with externals | Plain-language explanation of what the tag measures so external consumers understand the signal. |
| Display Range | Shared with externals | Optional numeric hints (Min, Max, Typical Value) for scaling charts and flagging out-of-range readings. |
| Std Dev | Shared with externals | Expected standard deviation for the measurement, useful for anomaly baselines. |
| Status | Lifecycle state | Moves through Draft → Active (Published) ↔ Paused → Archived. |
Two Ways to Create a Data Collection
Quick Create
Name your collection, pick the tags, and submit. The platform uses AI to generate aliases automatically. Best when you want to get started fast and refine later.
- Fastest path
- AI-generated aliases
- Edit aliases any time before publishing
Step-by-Step Wizard
A guided three-step flow: set the identity, select tags, then review and refine every alias, description, and display range before creating. Best for production-ready collections shared with external partners.
- Full control over every field
- AI aliases pre-filled and manually editable
- Inline display ranges and metadata
Who Sees What?
| Field | Owner | External Consumer | Notes |
|---|---|---|---|
Tag Name (e.g. FI-101) | Yes | No | Internal identifier — never exposed |
Alias (e.g. total_mill_feed_rate) | Yes | Yes | Public contract field |
| Description | Yes | Yes | Curated plain-language metadata |
| Min / Max / Typical Value | Yes | Yes | Optional display hints |
| Std Dev | Yes | Yes | Optional baseline hint |
| Collection Name | Yes | Yes | — |
| Live Telemetry Values | Yes | Yes (when Active) | Subject to share grants |
| Share Settings | Yes | No | Owner configures access and limits per share |
| API Keys | Yes | Yes | Owner creates direct keys; external recipients create share keys within owner limits |
How AI Alias Generation Works
When you create a Data Collection, the platform automatically generates a snake_case alias for each tag using a two-tier strategy:
- Short description (≤ 40 chars): Slugified directly with no AI call (for example,
"Total mill feed rate"→total_mill_feed_rate). - Long or missing description: An AI model generates a meaningful, readable alias from the tag name, description, and engineering units without leaking internal instrument codes.
All aliases are unique within the collection (collisions automatically receive _2, _3, etc.). You can manually edit any alias while the collection is editable.
Frequently Asked Questions
- Why can't I publish? It says tags need aliases.
Publishing requires that every tag has an assigned alias. Add or generate aliases for any missing tags and try again. - Do external consumers ever see real tag names like
FI-101?
Never. The external API only exposes curated aliases and collection-scoped tag UUIDs. - Can external consumers create their own API keys?
Yes. Collection owners create direct API keys for their own access. External share recipients create their own API keys bound to their share, within the max-keys and lifetime limits set by the owner.