# Case studies across engineering

The lectures carry their ideas on a few datasets chosen to run through the course: an
air-quality sensor network, a lab of wireless motes, a simulated chemical plant, concrete
mixes, and turbofan engines. The ideas themselves belong to no single field. A split that
leaks information inflates accuracy just as much for heartbeats as for engines. A seasonal
effect hides damage in a bridge the way drift hides a fault in a sensor.

Each page in this section takes one documented case from another branch of engineering and
reads it through the lectures. Every case rests on published work, every number on these
pages was checked against its source, and every page links back to the sections of the
notes it illustrates. They are optional reading. Pick the ones closest to your own field,
or the ones that cover an idea you found hard in class.

| Case study | Field | Connects to |
|---|---|---|
| [Heartbeat classification and the patient split](ecg-patient-splits.md) | Biomedical engineering | L9, L10, L12 |
| [Chest X-ray models that learned the hospital](chest-xray-shortcuts.md) | Biomedical imaging | L1, L10, L12 |
| [Validating a deployed sepsis alert](sepsis-alert-validation.md) | Health care analytics | L1, L10, L14 |
| [Forecasting electricity demand](electric-load-forecasting.md) | Energy systems and policy | L7, L8, L14 |
| [Near-perfect bearing fault classifiers on the CWRU benchmark](bearing-fault-benchmark.md) | Mechanical and electrical engineering | L7, L10, L12 |
| [Separating temperature from damage on the Z24 bridge](bridge-health-monitoring.md) | Civil and environmental engineering | L1, L6, L8 |
| [Searching 800,000 shape memory alloys with expected improvement](alloy-adaptive-design.md) | Materials science and engineering | L9, L14, L15 |