AutoGreen-MacBERT: Chinese Automotive Supply Chain ESG Sentences Annotated 2012-2024

Sentence-level text samples from annual, ESG, sustainability, and CSR reports of listed firms in China's automotive supply chain, covering 2012 to 2024. Each sentence is annotated for green strategic orientation, mixed expressions, and nine green action elements. The dataset was created by Chaohui Wang and is available under a CC-BY-4.0 license.

Use Cases

• Training Chinese NLP models for ESG text classification based on sentence-level annotations.
• Identifying green strategic orientation in corporate reports based on the substantive, symbolic, or non-green labels.
• Constructing firm-year TGSO (Total Green Strategic Orientation) indices based on the annotated action elements.
• Analyzing trends in green disclosures within the Chinese automotive supply chain over time.

Strengths

• Provides sentence-level annotations for three green orientation categories and nine specific green action elements.
• Covers a 13-year time range (2012-2024) from a specific industry and geographic context.
• Data is sourced from multiple official report types (annual, ESG, sustainability, CSR).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale training.

Provenance

Source

Chaohui Wang via figshare.

Collection Method

Collected from annual reports, ESG reports, sustainability reports, and CSR reports of listed firms.

Time Range

2012 to 2024

Freshness

Last updated 2026-05-28 00:50:57; freshness should be verified.

Geography

China