(Un)designing a Thermal Prediction System

Sensor- and weatherdata for a PhD research study into the implications of designing a thermal prediction system for shading and ventilating in domestic contexts. Part of the PhD thesis: "(Un)designing Learning Agents in Messy Domestic Contexts". This dataset includes sensor data (from the sensor kit that was developed for this project), API data (from weather stations; measurements and forecasts), and interface data (displayed output from the thermal prediction system; used for data visualizations of historic data and predicted temperatures). The sensor data was collected from four households living in Belgium and the Netherlands. The sensor kit consisted of four types of sensor devices that were installed in the households living areas. The sensor devices logged indoor and outdoor temperatures, sunlight intensity through selected windows, and window/curtain/door states. Temperature measurements from nearby weather-stations were accessed from an API called OpenWeatherMap. Global solar radiance measurements were accessed from an API called Meteoserver. Both API data streams were logged at hourly intervals. For the purpose of generating thermal predictions, APIs from the same platforms were used to access forecast data about outdoor temperatures and global solar radiance. The code underlying the thermal prediction system is available on Github: https://github.com/EmiliaViaene/Un-designing-a-Thermal-Prediction-System