A humble proposal for GeoAI: Epistemology, methodology, and relevance in curb ramp classification
Bibliographic record
Abstract
Abstract This paper critically examines how data availability and algorithmic constraints reshape research design, relevance, and outcomes in GIScience. Through a case study of automated curb ramp detection in Seattle, Washington, I demonstrate how the limitations of data, tools, and computational frameworks frequently shape the ambitions of early‐stage geographic inquiry. While GeoAI and machine learning offer new possibilities for spatial analysis, they also embed assumptions, value hierarchies, and technical limitations that influence what questions can be asked and what answers can be obtained. Using a random forest model and open‐source LiDAR and imagery data, I show how data sparsity, class imbalance, and aspatial training techniques complicate both accuracy and utility. Drawing on visual methods and explainable AI, I interrogate how the algorithm “learned” patterns and where it failed, revealing that AI‐identified relevance often diverges from socially meaningful goals. I argue for a reflexive approach to GeoAI—one that embraces error, foregrounds relevance, and resists the allure of algorithmic objectivity. The paper ultimately calls for centering research relevance alongside reproducibility and replicability in spatial data science, advocating for humility in the face of technological complexity .
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 | 0.133 | 0.199 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.010 | 0.006 |
| Science and technology studies | 0.008 | 0.138 |
| Scholarly communication | 0.019 | 0.037 |
| Open science | 0.006 | 0.015 |
| Research integrity | 0.008 | 0.012 |
| Insufficient payload (model declined to judge) | 0.005 | 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".