A critical review of use cases and insights from a large dataset of smart thermostats
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Residential buildings consume a significant portion (17 % in 2023) of the global primary energy. Smart thermostat has become a proven technology in the residential building sector that offers insights into energy efficiency, HVAC system operation, and indoor thermal comfort of occupants. Although there are an increasing number of studies using the available large scale smart thermostat dataset, there lacks a holistic review of the existing literature to understand what applications have been conducted and what outcomes have been offered. This paper reviews 57 articles published between January 2015 and March 2025 using the open access ecobee Donate Your Data (DYD) dataset, where >200,000 customers participated in the voluntary data donation program. Articles are analyzed by major application areas including occupant behavior and IEQ assessment, energy performance evaluation, HVAC operations and controls, and building thermal dynamics. Two major limitations of the DYD dataset are the lack of measured energy use of HVAC systems and the coarse city-level building location information and limits applications requiring energy use data and introduces errors in ignoring the urban microclimate effects influencing a home’s operation and performance. Gaps and challenges of using the ecobee thermostat dataset for research were analyzed. Future efforts should focus on improving data collection and fusing other datasets with the ecobee DYD dataset to unlock new applications and improve analytics accuracy. Furthermore, AI emerges as a powerful tool to help clean up, integrate, and analyze the thermostat dataset, create and calibrate energy models, as well as inferring residential building operation and performance at scale.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it