- ESTIMATION—Assessing the values and distributions of uncertain variables:
- Where is the communications satellite right now?
- How far is the user from the nearest branch when they deposit a check via a mobile app?
- NATURAL LANGUAGE PROCESSING—Giving context to unstructured text
- Is the term "woot" positive or negative the way it is used in a particular sentence?
- Is the term "crash" being used as a noun or a verb?
ANALYTICS: - RANKING—Ordering items by preference:
- Which battery vendor do I go to first, second, third?
- Which users should get a “try this new feature” offer?
- CHARACTERIZATION—Drawing conclusions from data:
- Why did the brake system fail?
- Why are users uninstalling the app?
- CLUSTERING—Grouping items by similarity/discovering similarities:
- Do different knee replacement patients fall into certain recovery classes?
- Do users with similar app-interaction patterns also share demographic patterns?
- RELATIONSHIP DISCOVERY—Finding correlations and cause-effect relations:
- Did the recent feature release increase the user interaction time with the app?
- What is the relationship between annual rainfall and bushels of soybeans per acre?
- VISUALIZATION—Exposing the salient aspects of complex data in simple pictures:
- How do we show the key features of networks that have a half million links?
- How do we picture the uncertainty in the aggregate data about user attrition?
CONTROL: - PLANNING AND SCHEDULING—Determining the best actions to take:
- How to distribute trolley cars to get tourists to the big comic book convention?
- What steps to take in what order and at what times to minimize user attrition?
- OPTIMIZATION—Determining the best choice among a set of possibilities:
- How do we maximize satisfaction with the user experience?
- What fuel mixture minimizes the overall cost of fleet vehicle operation?
S3 Data Science, copyright 2015. |
|