Notes

Writing system

Argument-driven notes on clinical AI, public datasets, research workflows, and systems analytics. The goal is to make the boundary of a claim easier to inspect.

Dataset autopsies Boundary notes Paper reality checks

Dataset Autopsies What does the data really represent? Public datasets, codebooks, survey design, leakage, missingness, and the small choices that shape model claims. Boundary Notes What can this method honestly say? Short clinical AI notes on prediction timing, SHAP, calibration, text boundaries, and review-ready interpretation. Systems / Water / Energy Notes What system boundary comes before the model? Environmental systems, water quality, energy analytics, LCA, sensing, and measurement limits in applied ML.

NHAMCS-ED: The Hard Part of Clinical AI Is Not the Model

NHAMCS-ED looks almost ideal for clinical AI experiments: it combines structured emergency department visit variables with short reason-for-visit text. But it is a national probability survey, not a hospital EHR extract. Across my IV fluid utilization and hospital admission prediction projects, the biggest lesson was that the hard part is often not the model itself; it is deciding what the model is allowed to know, what the data actually represent, and what claims the analysis can support. ...

SHAP Explains the Model, Not the Patient

SHAP Explains the Model, Not the Patient In clinical AI papers, I often see SHAP plots used as if they explain the disease process itself. The figure may look technical and persuasive, but the interpretation often goes too far: “This variable has a high SHAP value, so it is an important clinical cause.” That is usually not justified. SHAP is useful. It can show what a trained model relied on when making a prediction. But its boundary should be clear: ...

Beyond Excel: How Machine Learning is Revolutionizing Life Cycle Assessment

We all know the problem with Life Cycle Assessment (LCA): It’s too hard. To calculate the carbon footprint of a single product, analysts spend months digging through PDFs, manually entering data into spreadsheets, and struggling with missing values. In my recent paper published in PLOS Climate, I explored how we can modernize this outdated workflow using Data Science. The New Framework Instead of treating Machine Learning (ML) as a “black box,” we mapped specific ML techniques to the four standard ISO phases of LCA. ...