Identifying opportunities for statistical regression
Typical statistical analysis efforts which often become official statistical projects, start out with determining an objective and then, ultimately, determining the right approach to meet that objective.
Popular data science opinion declares determining an objective as establishing the purpose of a statistical analysis effort, then splits the purpose into three areas:
- Summarizing data (also called building a data profile)
- Exposing and exploring relationships between variables in the data
- Testing the significance of differences (between variables or groups within the data)
Summarizing data
If your statistical objective is to summarize data, you generate descriptive statistics, such as mean, standard deviation, variances, and so on.
Exploring relationships
If your statistical objective is to look for and learn about relationships in your data, you first examine your data for a form or, in other words, ask the question: does your data revolve around...