Making Sense of a Hot Mess: Cleaning and Validating Messy Administrative Data to study Supportive Housing in Winnipeg, Manitoba
Bibliographic record
Abstract
IntroductionWhile supportive housing (SH) is an important alternate to nursing home (NH) use, these data have never been linked to administrative records in Manitoba. By conducting linkages to other administrative records, we describe a process for cleaning and validating SH data, in preparation to conduct policy-relevant research. Objectives and ApproachSH data (N=516 units) from Winnipeg were received at the Manitoba Centre for Health Policy (MCHP) in three different files. File 1 (2004-2008; 1005 records) contained monthly client snapshots. File 2 (2008-2010; 1336 records) contained application, move-in, cancellation, and move-out dates. File 3 (2010-2011; 729 records) contained one line of text for each record showing the application, processing, and move-in/cancellation date. We used overlapping data from these files plus linkages to other data sources (Manitoba Population Registry, nursing home data, and Vital Statistics) to clean and assess the accuracy of SH data. ResultsThe original files contained 2039 people with 3070 records. From this we excluded: i) 215 records with unusable Personal Health Identification Numbers; ii) 949 records with missing SH move-in dates; iii) 691 records that did not match to the Manitoba Health Registry; and iv) 25 records where data did not match to the NH, hospital, or Vital Statistics files. The result was 1190 people each with one record. SH move-out dates were often missing from these records. This field was imputed from other data sources (NH, Vital Statistics). Some people transferred between SH sites, and these data were retained in the same record. Aside from the first year of operation when capacity was low, most SH dwellings operated at 80-100% occupancy annually. Conclusion/ImplicationsUsing several verification methods including linkages to other data sources, we successfully cleaned and verified the accuracy of the SH data for use at MCHP. High annual SH occupancy rates suggest that the file contains the vast majority of SH users, and can now be used in follow-up research.
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 | 0.014 | 0.050 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.008 |
| Science and technology studies | 0.005 | 0.002 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".