---
license: cdla-permissive-2.0
tags:
- energy
- smart meter
- Centre For Net Zero
size_categories:
- 1M<n<10M
---


This dataset contains 10 million synthetic load profiles of trained on over 1 billion smart meter readings from 1 million Octopus Energy UK households sampled between 1st March 2024 and 1 March 2025. 
The dataset comprises of two sub populations:

1. Households with any form of LCTs (heat pumps, EVs, Solar PVs etc): The distribution here is reflective of Octopus's customer base
2. Households with no LCTs (all LCT columns = False): Octopus's non-LCT households were resampled to create a UK representative sample.

**The smart meter profiles are conditioned on labels such as the:**

- `Property types`: house, flat, terraced, detached, semi-detached etc
- `Energy performance certificate (EPC) rating`: A/B/C, D/E, F/G etc
- `Low Carbon Technology (LCT) ownership`: heat pumps, electric vehicles, solar PVs etc
- `Seasonality`: weekday vs weekend, month of year
- `Tariff types`: standard, smart, automated, economy 7
- `Location Cluster`: An unsupervised learning approach is used to group regions on the UK ([LSOAs](https://www.ons.gov.uk/methodology/geography/ukgeographies/censusgeographies/census2021geographies)) into 30 clusters. This model is trained using features related to the energy consumption profiles of households in each area. The cluster label is included in our dataset and we have also uploaded a mapping from LSOA -> cluster. Using this mapping, you can filter synthetic profiles for specific regions of the UK. 


For more information about Faraday and our method to generate synthetic smart meter profile, please refer to the [workshop paper](https://www.climatechange.ai/papers/iclr2024/43) that Centre for Net Zero presented at ICLR 2024. For more information about OpenSynth, please visit our Github repository https://github.com/OpenSynth-energy/OpenSynth. For more news and updates on OpenSynth, please subscribe to our mailing list here.