File size: 7,020 Bytes
2020a6f c309992 2020a6f c309992 ba796d2 c309992 ba796d2 c309992 ba796d2 c309992 ba796d2 c309992 ba796d2 c309992 ba796d2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 |
---
license: pddl
tags:
- climate
size_categories:
- n<1K
dataset_info:
features:
- name: scope_1
dtype: double
- name: scope_2_market
dtype: double
- name: scope_2_location
dtype: double
- name: scope_3
dtype: double
language:
- en
---
# Dataset Card for Dataset Name
A dataset of 100 corporate sustainability reports with manually extracted scope 1, 2 and 3 greenhouse gas emission values.
## Dataset Details
### Dataset Description
Data about corporate greenhouse gas emissions is usually published only as part of sustainability report PDF's, which is not a machine-readable format. Interested actors have to manually extract emission data from these reports, which is a tedious and time-consuming process. An automatic information-extraction system could solve this issue.
To evaluate such information-extraction systems and to encourage research into solving this task, a dataset of sustainability reports and manually-extracted emission values is created and published.
While this dataset is intended for evaluation, the accompanying [sustainability-report-emissions](https://huggingface.co/datasets/nopperl/sustainability-report-emissions) dataset is intended for training/finetuning models.
- **License:** Open Data Commons Public Domain Dedication and License (PDDL)
### Dataset Sources [optional]
- **Repository:** https://github.com/nopperl/corporate_emission_reports
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
The dataset is intended to be used to evaluate automatic systems to extract machine-readable greenhouse gas emission data from sustainability reports.
## Dataset Structure
- `id` (string): unique instance id, ex. `"0012"`.
- `emission_year` (int): the year of the extracted emissions, which is useful in reports which contain information about multiple years.
- `scope_1` (double): total scope 1 emissions in metric tonnes of CO2eq.
- `scope_2_market` (double): total market-based scope 2 emissions in metric tonnes of CO2eq.
- `scope_2_location` (double): total location-based scope 2 emissions in metric tonnes of CO2eq.
- `scope_3` (double): total scope 3 emissions in metric tonnes of CO2eq.
- `scope_1_page` (list<int>): set of pages containing total scope 1 emission data.
- `scope_2_market_page` (list<int>): set of pages containing total market-based scope 2 emission data.
- `scope_2_location_page` (list<int>): set of pages containing total location-based scope 2 emission data.
- `scope_3_page` (list<int>): set of pages containing total scope 1 emission data.
- `url` (string): the URL to the sustainability report PDF.
- `sha256` (string): SHA-256 hash string of the report PDF to ensure the integrity of downloaded files.
- `subset` (string): indication of whether the report comes from the set of Euro Stoxx 50 (`eurostoxx`), NYSE (`nyse`) or Nikkei 225 (`tyo`) corporations.
The remaining 15 fields contain the data for each of the [15 scope 3 emission categories](https://ghgprotocol.org/scope-3-calculation-guidance-2).
The dataset only contains the URL to the report PDF. A helper script to download these files is provided at: https://github.com/nopperl/corporate_emission_reports/blob/main/download_documents.py.
## Dataset Creation
### Curation Rationale
To our knowledge, there is no publicly-available dataset containing manually extracted (self-reported) greenhouse gas emission data from sustainability reports. Hence, this dataset was collected to enable the evaluation of automatic information-extraction systems.
### Source Data
The dataset is based on sustainability reports from corporations in Europe, North America and Asia.
#### Data Collection and Processing
To ensure geographic diversity, the sustainability reports are sourced from three sets of corporations. The first set consists of 39 corporations tracked by the Euro Stoxx 50 stock index as of 18 September 2023. Note that out of the missing 11 corporations of Euro Stoxx 50, 5 are not considered as no publicly-available sustainability reports were found and 6 are not considered it was not possible to extract unambiguous emission values. The second set is a random selection of 39 corporations listed on the New York Stock Exchange as of December 2023. The third set is a random selection of 22 corporations tracked by the Nikkei 225 index as of October 2023. Note, that this selection strategy is intentionally biased towards larger corporations due to the assumption that they have a higher likelihood of publishing sustainability reports.
For every corporation, the most recent sustainability report is downloaded from the official source. In some cases, the sustainability report is part of a larger annual report.
Sustainability reports based on the [Carbon Disclosure Project](https://www.cdp.net/en/guidance) or [Global Reporting Initiative](https://www.globalreporting.org/standards/) templates are not considered as they already follow a consistent structure.
#### Who are the source data producers?
The sustainability reports are produced by corporations themselves and optionally verified by third parties. Thus, they only contain self-reported emission information.
### Annotations [optional]
The sustainability reports are annotated with manually-extracted emission data, which forms the main purpose of this dataset.
#### Annotation process
The annotation was based on the greenhouse gas emission definitions of the [GHG Protocol Corporate Standard](https://ghgprotocol.org/corporate-standard):
- Scope 1: A reporting organization’s direct GHG emissions.
- Scope 2: A reporting organization’s emissions associated with the generation of electricity, heating/cooling, or steam purchased for own consumption.
- Scope 3: A reporting organization’s indirect emissions other than those covered in scope 2.
Only emission data about the reporting corporation was extracted, invidiual values for subsidiaries were ignored.
Scope 2 emissions are annoted as market-based by default if no indication about the calculation method is given in the report.
Values which could not be unambiguously extracted were noted as missing.
No automatic tools were used in the extraction process.
The extracted data is not validated by a third party or verified against other data sources.
#### Who are the annotators?
The data was annotated by a single person without special expertise in sustainability reporting.
#### Personal and Sensitive Information
The dataset contains only public information.
## Bias, Risks, and Limitations
The emission information was extracted by a single non-expert. No guarantee can be given that the data is completely correct.
The dataset does not contain sustainability reports of small enterprises or non-profit organisations.
Even though some care was taken to ensure geographic diversity, the dataset does not include sustainability reports from large parts of the world.
## Citation [optional]
**BibTeX:**
[More Information Needed]
|