Room Occupancy Estimation with Seven Wireless Sensor Nodes

Overview

The dataset uses seven wireless sensor nodes to estimate room occupancy from temperature, light, sound, CO₂ and PIR signals. The target Room_Occupancy_Count ranges from zero to three.

Measured structure

The CSV has 10,129 data rows and 19 columns: Date, Time, four temperature channels, four light channels, four sound channels, CO₂, CO₂ slope, two PIR channels and the occupancy target. Thus 18 input fields include the two date/time fields; they are not 18 separate physical sensor channels. No empty cells were found.

Timing

The source describes approximately 30-second sampling. Although its collection narrative refers to four days, the released CSV contains seven distinct calendar dates: 22–26 December 2017 and 10–11 January 2018. Do not describe this file as four consecutive days. The citation uses the publisher’s 2018 citation year; the later UCI donation date is 2023.

Source and provenance

Dataset DOI: https://doi.org/10.24432/C5P605.