When you think of the purpose of Data Science or Machine Learning, what answers come to mind? On the basic level, Data Analysts and Data Scientists use Analytics and Machine Learning Algorithms to solve business problems. But, what does that really mean?
In the book, Data Science for Business by Foster Provost & Tom Fawcett, they provide great examples of how Data Science is used in multiple industries. Here are four common ones:
1. Classification & Probability Estimation
Helps answer questions such as: Which customers will likely respond to a given offer? It attempts to predict the probability that an individual customer will respond to the offer or not. And the two classes would be: will respond and not respond.
2. Regression
Helps answer questions such as: How much will a customer use this service? Regression is a supervised learning algorithm that estimates a constant value for each individual based on historical data.
3. Clustering
Groups individuals within a population based on similarities. While data points, or customers within each group will share similarities, there are distinctions between each group. This is an unsupervised algorithm as there is no target value and no guarantee that the similarities will always be insightful.
4. Co-occurrence Grouping
Helps answer questions such as: Which items are purchased together? Another name is market-basket analysis which attempts to find associations between items based on transactions involving them. For example, based on data it may show when people buy a web camera, they also buy a microphone. This could result in recommending to the business that they should run a special promotion or design a specific product display to encourage more people to buy both items together.
Not every problem needs a Machine Learning solution, but these are just a few ways Data Scientists and other Data professionals today use Machine Learning to solve business problems.
