A 2026 study by Abiola Akanmu presents a taxonomy of decision-relevant insights for construction visual analytics. The dataset is derived from the final round of a Delphi process involving construction professionals, analyzing 128 insight-context items across eight scenarios including safety, progress, and quality control. It provides a practitioner-grounded framework for developing image-based decision-support tools and training.
Use Cases
- Designing construction dashboard interfaces based on the identified cross-cutting insight categories like location-based and temporal information.
- Developing workforce training for visual data interpretation based on the consensus insights for safety, progress, and inspection scenarios.
- Prioritizing features for image-based decision-support tools based on the high-, mid-, and lower-rated applicability bands of insights.
- Informing computer vision research objectives by focusing on extracting practitioner-endorsed information types such as dimensional and condition/specification data.
Strengths
- Dataset is based on 128 insight-context items analyzed from a structured Delphi process with construction professionals.
- All items achieved expert consensus under the median absolute deviation criterion, indicating a validated foundation.
- Results differentiate insight applicability across high-, mid-, and lower-rated bands, providing a nuanced view of practical importance.
- Released under a permissive CC0-1.0 license, allowing for unrestricted use and redistribution.
Limitations
- Row count is unknown, which may limit suitability assessment for large-scale modeling.
- Column-level documentation is absent; field semantics must be inferred after download from the provided CSV or RTF files.
- The dataset is very small at 20.5 KB, indicating limited scope and likely containing summary or coded results rather than raw survey data.
Provenance
- Source
- figshare, authored by Abiola Akanmu.
- Collection Method
- Derived from the final round of a Delphi study involving construction professionals.
- Freshness
- Last updated 2026-05-28 15:09:25; freshness should be verified.