Electricity: Subsampled Data for Energy Prediction with 10 Classes

A 2,000-row subsample of an electricity consumption dataset, derived from the OpenML dataset with ID 44156. It was generated by Eddie Bergman using a script that selected 100 columns and 10 target classes. The data is provided under a US public domain license.

Use Cases

• Training classification models to predict electricity consumption patterns based on the 100 available features.
• Benchmarking model performance on a controlled, stratified subset of energy data.
• Analyzing feature importance for energy prediction within a reduced feature space of 100 columns.
• Prototyping algorithms for multi-class (10-class) prediction on tabular energy data.

Strengths

• Subsampling parameters are explicitly documented: seed=0, nrows=2000, ncols=100, nclasses=10.
• The generation script indicates stratified sampling was used to preserve class distribution.
• Derived from a known source dataset (OpenML ID 44156), providing traceability.

Limitations

• Row count is unknown for the original source dataset, limiting assessment of the subsample's representativeness.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

OpenML dataset ID 44156 (electricity).

Collection Method

Algorithmic subsampling via a Python script provided in the description.