Office plug-load dataset for non-intrusive load monitoring in buildings using an open metering platform

This repository contains the UALM6 office plug-load dataset associated with the manuscript entitled “Office plug-load dataset for non-intrusive load monitoring in buildings using an open metering platform”. The dataset was acquired using Open Multi Power Meter v2 (OMPM v2), an open, low-cost, multi-channel metering platform designed for synchronized aggregate and appliance-level electrical monitoring in office and institutional buildings. The monitored scenario includes one aggregate plug-load channel and five individual office devices: a desktop workstation, a laser printer, a smart display, a laptop computer, and an LED desk lamp. The acquisition campaign was carried out under natural office operating conditions, without programmed switching sequences or artificial load automation. This preserves realistic standby periods, sparse events, quasi-continuous low-power operation, variable device usage, and natural interactions among office plug loads. The dataset covers a continuous period from 5 June 2026 at 11:13:00 to 15 June 2026 at 11:40:00, corresponding to approximately 10 days and 27 min of operation. Six electrical meters were sampled with a nominal period of 15 s, equivalent to a sampling frequency of 0.0667 Hz. Each meter contains 57,708–57,709 samples, depending on the channel, resulting in a total of 346,252 time-stamped records. For each timestamp, the following variables are provided: voltage, current, active power, frequency, power factor, ambient temperature, and relative humidity. The time series are highly regular, with a constant 15 s interval between consecutive samples and no duplicated timestamps or internal gaps detected. The repository includes the raw CSV measurements, the NILMTK-compatible HDF5 dataset, metadata files, and scripts required to reproduce the dataset conversion and baseline non-intrusive load monitoring (NILM) evaluation. The dataset is intended to support reproducible research on NILM, office plug-load monitoring, load observability, low-power device detectability, and building energy management in non-residential environments.