Benchmarking operational performance of buildings by text mining tenant surveys
Bibliographic record
Abstract
Facility managers of large commercial and institutional buildings periodically collect text-based survey data from their tenants. While these large and amorphous datasets contain valuable information to benchmark operational performance and identify anomalies, it is time and resource-intensive to hire employees to read and analyze the datasets and extract insightful information from them. This paper presents a natural language processing-based methodology to extract operational insights from tenant survey databases. It also incorporates the verification of extracted complaint patterns using computerized maintenance management systems (CMMS) on a smaller scale. Tenant survey databases are comprised of free-text responses from tenants regarding annual/bi-annual survey responses that building managers request as a source of solicited feedback. CMMS databases consist of unsolicited complaints that are logged by tenants who are under discomfort/dissatisfaction with no additional prompt from a building operator/manager. The effectiveness of this methodology is demonstrated by gaining operational insights from tenant feedback gathered using survey data from a large office building in Ottawa, Canada. Different algorithms for sentiment analysis, association rule mining, and topic modeling are employed in the analysis to consolidate the textual data into common thermal and maintenance complaint categories. The accuracy of different text analytics algorithms is compared, and their effectiveness in analyzing tenant survey responses is discussed. Patterns of unsolicited tenant work order requests are contrasted to those observed in the survey responses. The results indicate that the floors that frequently submit work order requests are also likely to submit a large number of negative survey responses.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.014 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".