This document discusses approaches for annotating heterogeneous data, with a focus on applications in transportation domains. It defines annotation as tagging or labeling data with metadata. Heterogeneous data comes from different sources and formats and at different granularities. Annotation can help with data integration, search, and addressing issues from diverse data schemas. The document reviews manual, semi-automated, and automated annotation techniques, and provides examples of rule-based and training dataset driven annotation. It also discusses using annotation for traffic data analysis like time estimation and accident avoidance. Overall, the document provides an overview of heterogeneous data annotation with a transportation domain application focus.