A Home Assistant-Based Residential HEMS Dataset for HVAC Control, Shiftable Appliance Scheduling, PV Generation, and Real-Time Pricing

This dataset provides a simulated residential home energy management system (HEMS) data environment based on Home Assistant household automation and synchronized online exogenous data. It was developed to support research on HVAC control, thermal comfort, shiftable appliance scheduling, photovoltaic generation, real-time electricity pricing, and residential demand-side energy management. The household-related variables were generated using a Home Assistant-based scenario representing a four-person household. The simulated variables include occupancy count, washing machine and dishwasher requests, appliance energy profiles, desired indoor temperature, active comfort indicator, base load, HVAC operating state, and HVAC load. These data were combined with online exogenous signals, including CAISO real-time electricity prices, NSRDB-based photovoltaic generation, and outdoor temperature for the Los Angeles / SP-15 California region. The dataset is organized at a 30-minute time resolution and covers the period from 2020-01-01 00:00:00 to 2022-09-07 23:30:00. It includes a full dataset file and predefined train, evaluation, real-test, and comprehensive-test splits to support reproducible model development and evaluation. This dataset is not an optimized controller or an intelligent HEMS by itself; rather, it provides a reusable simulation-based data environment for researchers to implement, train, test, and compare their own control, optimization, and machine-learning methods for residential HEMS applications.