How do Data Scientists approach predictive modeling?

Predictive modeling by the Data Scientist requires feature selection: selection of a relevant set of features in the data on which an outcome may depend. This is followed up by the choice of an appropriate algorithm, which can be something like linear regression or decision trees, to make up the model. The model then is applied to make predictions on new data, after being properly trained on historical data and, on top of that, validated using a variety of techniques, such as cross-validation.