ontology-mapper_skill

Overview

This skill maps construction data fields to standard ontologies (IFC, COBie, Uniclass, OmniClass, MasterFormat, UniFormat) to enable semantic interoperability across systems. It automates semantic matching, suggests relation types (equivalent, broader, narrower, related), and returns confidence levels for each mapping. The tool is designed for data ingestion, export normalization, and integration workflows in AEC data pipelines.

How this skill works

The mapper loads simplified ontology concept sets and predefined rules, then normalizes input field names and values for matching. It uses exact, partial, and synonym-based similarity checks to propose semantic mappings and assigns a confidence tier (exact, high, medium, low, uncertain). Results include multiple candidate mappings, the best match, and an overall mapping status for each field.

When to use it

• Normalizing incoming construction datasets before BIM/asset import
• Mapping legacy CSV/property sets to IFC or COBie for handover
• Generating semantic mapping reports for integration projects
• Reconciling disparate classification systems (Uniclass, OmniClass, MasterFormat)
• Automating data validation and coverage checks during model exchange

Best practices

• Provide known source ontology when available to improve matching accuracy
• Normalize your domain-specific synonyms before bulk mapping to increase exact matches
• Review mappings with medium/low confidence and record manual corrections back into mapping rules
• Run mapping on representative sample data first to tune synonyms and thresholds
• Export mapping reports with coverage and confidence distribution for stakeholder review

Example use cases

• Map a proprietary door type field to IFC IfcDoor and Uniclass/ MasterFormat codes for procurement and cost estimating
• Convert a set of facility CSV attributes to COBie spaces and components for FM handover
• Create a reconciliation report showing unmapped fields and suggested ontology concepts for data clean-up
• Automate pre-import mapping that tags incoming property sets with ontology IDs and confidence for downstream validation