Edge‑Based Dataset of Irradiance and Temperature on Urban Building Roof and Facade Surfaces
Abstract
This dataset contains high‑resolution measurements of solar irradiance and ambient temperature collected on roof (horizontal plane, I1/T1) and facade (vertical plane, I2/T2) surfaces of an urban university building in Medellín, Colombia. The monitoring system was implemented using an edge computing device that enabled local acquisition, validation, and transmission of sensor data to the cloud in real time. The dataset is organized as time series in both CSV and Parquet formats, ensuring accessibility and reproducibility. It provides continuous records from January 30, 2026 (14:30 local time) to June 12, 2026 (12:39 local time), with an average sampling interval of approximately 1.71 seconds. Measurements on dual orientations (roof vs. façade) offer valuable insight into the variability of solar exposure in building‑integrated photovoltaic (BIPV) contexts, highlighting the influence of surface inclination and surrounding urban elements. The dataset underwent a structured processing workflow to ensure quality and consistency. Initially, the consolidated CSV file was loaded into the edge computing environment, and only relevant columns were selected (sensor ID, measurement value, and units). Timestamps were converted from UTC to local time in Medellín (UTC‑5). Subsequently, null records were removed, and out‑of‑range values were excluded, specifically irradiance and temperature values below zero. Files referring to irradiance measurements are denoted with the prefix irr_, while files referring to temperature measurements are denoted with the prefix temp_. The suffix 1 corresponds to sensors installed on the roof (horizontal plane), and the suffix 2 corresponds to sensors installed on the façade (vertical plane). Parque i ├── Horizontal measurement (I1, T1) - roof │ ├── irr_1.csv │ ├── irr_1.parquet │ ├── temp_1.csv │ └── temp_1.parquet └── Vertical measurement (I2, T2) - façade ├── irr_2.csv ├── irr_2.parquet ├── temp_2.csv └── temp_2.parquet
