Table 2 Chemical Engineering Construction Data Type Classification.

From: Multi-source heterogeneous data fusion and intelligent prediction modeling for chemical engineering construction projects based on improved transformer architecture

Data Type

Data Source

Data Characteristics

Processing Method

Structured Numerical

Sensor Networks, SCADA Systems

High-frequency time series, precise measurements, standardized formats

Normalization, outlier detection, temporal alignment

Structured Categorical

Project Management Systems

Discrete labels, hierarchical classifications, workflow states

One-hot encoding, ordinal encoding, embedding layers

Semi-structured Temporal

Equipment Logs, Maintenance Records

Irregular timestamps, mixed data types, nested structures

Schema extraction, temporal interpolation, missing value imputation

Semi-structured Spatial

Geographic Information Systems

Coordinate systems, spatial relationships, multi-scale representations

Coordinate transformation, spatial indexing, resolution harmonization

Unstructured Textual

Technical Documentation, Reports

Natural language, domain-specific terminology, variable length

Text preprocessing, tokenization, semantic embedding

Unstructured Visual

Engineering Drawings, Inspection Images

High-dimensional pixel data, geometric features, annotation layers

Image preprocessing, feature extraction, object detection

Back to article page