>
Hoxton Data
Human accuracy at machine speed
About
We build validated data solutions. Our demo tool automatically extracts specific impact metrics from corporate annual reports. These solutions can be used to source verified datapoints for decision-making (Track 1), to validate and improve the quality of analysts’ or researchers’ work through the tool (Track 2), and to create high-quality proprietary datasets that fuel internal ML/AI models (Track 3). Our prototype is trained on a manually-validated FTSE 100 annual report dataset for 2023/24 that achieved c.90% weighted accuracy on the first iteration – which was then validated by human experts to reach full accuracy. Commercially we offer three tracks (1) deliver cleaned and verified impact or financial data points on request, (2) license & tailor the extraction-validation tool, or (3) run expert workshops to capture processes and build bespoke proprietary internal datasets that clients can embed into their own ML/AI workflows – the latter being ideal where high-quality, custom training data is the most valuable asset.
| Value | Unit | Metric | Year | Source document | Page |
|---|---|---|---|---|---|
| 2181890.0 | tCO2e | Scope 1 emissions | 2020 | SMDS Annual Report 2024 | 77 |
| 100.7 | Mn MWH | Energy consumption | 2023 | IAG Annual Report 2023 | 85 |
| 6.53 | MTCO2e | Scope 3 emissions | 2023 | IAG Annual Report 2023 | 85 |
| 5211.0 | tCO2e | Scope 1 emissions | 2022 | REL Annual Report 2023 | 42 |
| 8207.0 | tco2e | Scope 3 emissions | 2021 | PHNX Annual Report 2022 | 24 |
| 4317.0 | tCO2e | Scope 1 emissions | 2023 | REL Annual Report 2023 | 42 |
| 140.0 | tCO2e | Scope 1 emissions | 2023 | III Annual Report 2024 | 68 |
| 9.27 | MtCO2e | Scope 1 emissions | 2024 | SSE Annual Report 2024 | 107 |
| 5.48 | MTCO2e | Scope 3 emissions | 2022 | IAG Annual Report 2023 | 85 |
| 104.5 | MtCO2e | Scope 3 emissions | 2022 | AAL Annual Report 2023 | 139 |
| ... | ... | ... | ... | ... | ... |
Data
Our initial dataset focuses on FTSE 100 annual reports (2023–2024), extracting greenhouse gas (GHG) emissions across Scope 1, Scope 2 (market- and location-based), and Scope 3, along with energy consumption metrics. Each datapoint is automatically identified, then validated by human experts to ensure accuracy and reliability. This foundation dataset will be further refined, expanded, and made openly accessible as part of our commitment to transparency.