Question
A data scientist needs to:
- Build a predictive model that gives the likelihood that a car will get a flat tire.
- Provide a data set of cars that had flat tires and cars that did not.
All the cars in the data set had sensors taking weekly measurements of tire pressure similar to the sensors that will be installed in the cars consumers drive.
Which of the following is the most immediate data concern?
- A. Granularity misalignment Correct
- B. Multivariate outliers
- C. Insufficient domain expertise
- D. Lagged observations
Explanation
A flat tire is an event that happens at a specific moment, while the tire pressure is only measured once a week, so the predictor and the outcome are recorded at different granularities. A pressure drop that leads to a flat can happen entirely between two weekly readings, which makes the measurements too coarse to predict the event; this granularity mismatch is the first thing to resolve. Lagged observations are a modeling technique for time-series features rather than the immediate problem here, and nothing in the scenario points to outliers or a lack of domain expertise.
