Dataset

TUDelft-Electricity-Consumption-1.0

TUDelft-Electricity-Consumption-1.0 is a high-resolution, open-source time series dataset of household electricity load published on Hugging Face by OpenSynth. It draws on multi-country trials covering thousands of households under different tariff regimes and is distributed in Parquet format.

Open on Hugging Face

Data preview

Subset

Loading preview...

Dataset card

What this dataset contains

The dataset holds household-level electricity consumption measurements at sub-hourly resolution. According to the description, it includes a 2013 Dynamic Time-of-Use tariff trial with around 1,100 Dutch households exposed to variable pricing (high, low, normal) plus a flat-rate cohort of roughly 4,500 households, and the source materials also reference high-resolution records from multiple countries. Price schedules accompany the consumption series. The train split alone contains 774,794,036 rows, placing the size in the 100M–1B range.

Structure and fields

Records are organized as a single tabular Parquet file with four columns: id (string identifying the household), datetime (timestamp[ns]), target (float64 normalized load), and category (string denoting the sampling interval such as 15m, 30m or 60m). Example loading is performed via dask.dataframe.read_parquet on the path hf://datasets/Weijie1996/load_timeseries/30m_resolution_ge/ge_30m.parquet. The schema supports straightforward conversion to pandas, polars or other DataFrame libraries.

How analysts use it

The repository documents use in research on transformer-based and flow-based generative models for residential load profiles and few-shot learning across thousands of domains. For investment analysts, the rich, granular household-load series combined with tariff information could plausibly inform energy-utility demand modeling, smart-meter vendor benchmarking, and consumer-energy-product revenue forecasting, although the dataset is academic in origin and has a permissive CDLA license tagged on the hub.

Documentation & references 1