ontology-mapper

What This Skill Does

The ontology-mapper is a specialized OpenClaw agent skill designed to bridge the gap between heterogeneous construction data schemas. Based on the DDC (Data-Driven Construction) methodology, it allows users to harmonize disparate data sources—such as BIM models, CAD exports, and legacy database records—by mapping them to standardized industry ontologies. Supported standards include IFC (Industry Foundation Classes), COBie, Uniclass, OmniClass, MasterFormat, and UniFormat.

The skill provides a structured framework for semantic interoperability. It evaluates source data points against defined ontology concepts, assigns relationship types (like equivalent, broader, or part-of), and calculates confidence intervals. This ensures that construction data remains machine-readable and consistent throughout the project lifecycle, reducing the risk of errors during project handover or collaborative multidisciplinary workflows.

Installation

To integrate the ontology-mapper into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/datadrivenconstruction/ontology-mapper

Ensure your agent has the necessary permissions to access your data directories and that you have initialized the OpenClaw CLI prior to installation.

Use Cases

• Project Handover: Automating the transition of facility management data from BIM models to COBie spreadsheets.
• Data Standardization: Aligning legacy asset management databases with current classification systems like Uniclass 2015.
• Interoperability Analysis: Detecting semantic gaps when migrating data between different CAD or BIM software versions.
• Automated Specification Mapping: Linking specific design elements to global master formats to streamline procurement and cost estimating.

Example Prompts

1. "Map the column headers in project_data_v2.csv to the standard IFC4 schema and provide a confidence report for all fields marked 'uncertain'."
2. "Analyze the current project hierarchy against UniFormat standards and flag any components that don't fit into the existing classification tree."
3. "Generate a mapping report comparing our internal BIM object names to OmniClass; list all unmapped fields and suggest potential matches for those with a confidence below 70%."

Tips & Limitations

• Confidence Thresholds: Always review mappings with a 'low' or 'uncertain' confidence level manually. The skill uses algorithmic inference which may require human expert validation for complex, custom architectural elements.
• Ontology Selection: Choose your target ontology carefully; IFC is best for geometric and spatial data, while COBie is superior for facility operations and asset maintenance.
• Data Quality: The effectiveness of the mapper is directly proportional to the cleanliness of your source data. Pre-processing datasets to remove duplicates or incomplete entries will yield significantly better semantic accuracy.
• Custom Mapping: If your organization uses proprietary classification systems, utilize the OntologyType.CUSTOM enum to extend the skill's capabilities to your unique organizational standards.