Types of Questions asked to Data Scientists

1. Descriptive - describe a set of data
• Descriptions cannot be generalized without adding statistical modelling
2. Exploratory - find relationships you didnt know about
• Exploratory analysis should not be alone used for generalizing/predicting
• Correlation doesnt imply causation
3. Inferential - use a relatively small sample of data to say something about the bigger population
• Inference is the common goal of statistical analysis
• Inference involves estimating both the quantity we care about and the certainty of that estimate
• Inference depends heavily on both the population and the sampling scheme
4. Predictive - To use the data in some objects to predict the data in other objects
• If X predicts Y then it doesnt mean that X causes Y
• More and more data works well with reasonable models
5. Causal - To find what happens to one variable when you change another variable
• Usually randomized variables are used for causation
• There are approaches to infering causation in non-randomized studies, but they are complicated and sensitive to assumptions
• Causal relationships are identified as average effects, but may not apply to every individual
• Causal models are usually the gold standard for data analysis
6. Mechanistic - physics