A daily CO2 emissions dataset for 14 countries and regions globally, spanning from 1970 to 2024. The dataset was constructed by Tao Li using over two million records of near-real-time electricity generation, traffic, and industrial output since 2019, with machine learning applied to simulate relationships between emissions and predictors. It encompasses four sectors: power, industry, residential, and transport.
Use Cases
- Modeling daily emissions trends based on the 1970-2024 time series.
- Analyzing sectoral contributions to CO2 output based on power, industry, residential, and transport categories.
- Training ML models for emissions prediction based on relationships with temperature and time surrogates.
- Comparing national and regional emissions trajectories based on coverage of 14 countries and regions.
Strengths
- Covers a long temporal range from 1970 to 2024.
- Includes data for four distinct sectors: power, industry, residential, and transport.
- Built from over two million records of near-real-time activity data.
- Employs machine learning to model nonlinear, time-varying relationships.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Geographic coverage is limited to 14 countries and regions.
Provenance
- Source
- Tao Li via figshare.
- Collection Method
- Constructed from collected activity records using machine learning to simulate emissions relationships.
- Time Range
- 1970 to 2024.
- Freshness
- Last updated 2026-04-14 07:02:08.
- Geography
- 14 countries and regions globally.