Practical Predictive Analytics

Using all of the data

However, many data scientists will choose to use the entire population as the basis of analysis, instead of a sample. I suspect one reason is that often it is difficult to obtain a reliable sample, especially when the data has many different sources, and some are of unknown data quality. Good reliable data costs money to collect. Another reason is that the data scientists may not be familiar with sampling techniques, or believe that more is always better. Therefore, it is important to understand some important points about using all of the data:

• You can use all of the data if it accurately represents the underlying population. If all the data is the underlying population, you can't do better than that, but proceed with ...