1. Descriptive analytics
Hi, I'm Carl and I'll be your second instructor for this course. In this chapter, we'll discuss the four types of analytics, starting with the basis: descriptive analytics.
2. Analytics overview
In this overview, you can see the four main types of analytics: descriptive, diagnostic, predictive, and prescriptive analytics. Each type of analytics has its own main focus.
The more we go to the right, towards prescriptive analytics, the more advanced questions we can answer. However, this does not mean that the most advanced analytics types are always the best choice.
For instance, choosing prescriptive analytics when descriptive analytics would suffice, needlessly complicates the analysis and increases the time it takes.
The suitable type of analytics depends on the question we want to solve. In the case of descriptive analytics, the leading question is: 'what is happening?'.
3. Why use descriptive analytics?
Descriptive analytics focuses on summarizing and visualizing the data.
This type of analytics can be used to get to know the data, based on its variables. Variables are the characteristics of interest we have measured or observed. For example, temperature in a weather dataset.
Descriptive analytics can also be used to investigate relationships or patterns in the data, for example, whether specific groupings occur or one variable is closely related to another.
Lastly, descriptive analytics can be used as a preparation step for more advanced techniques, like building a model.
4. Common techniques
In its most basic form, a common technique of descriptive analytics consists of the calculation of descriptive statistics, like averages and maximum or minimum values of variables. These help us answer questions like 'What is the most common value?' or 'How spread out are the values?'
Visualizing the data is also helpful to get to know the characteristics of the data, in particular to see patterns, groupings, or other relationships in the data.
Outlier detection helps to find values that are very different from the majority, which can help to find errors.
A more advanced and comprehensive form of descriptive analytics is exploratory data analysis, or EDA.
5. Exploratory data analysis
The focus of exploratory data analysis, or EDA, as the name implies, lies in exploring the data: assessing its main characteristics, finding relationships, patterns or groups, and suggesting hypotheses for future analysis.
Often it is contrasted with hypothesis testing, which is more formal and specific, while EDA looks at data in a broad, open way to discover something that is not known yet.
It combines different techniques to gain a comprehensive picture of the data, including basic descriptives and more advanced techniques such as cluster analysis to find groupings. It also places a strong emphasis on visualization to gain additional insights.
It can be used as preparation for further analysis and offer valuable insights on its own, for example, to discover data problems.
6. Case study: ice cream sales
Suppose you work for an ice cream company and are tasked with analyzing ice cream sales. You are particularly interested in which flavor sells best.
Using descriptive analytics, you could count the number of sales per flavor and, if needed, separate the numbers per store or month.
The insights gained from this can then be used to limit the number of flavors to the most popular or select new flavors to sell based on the characteristics of the already popular flavors.
7. Let's practice!
Now that you have learned about descriptive analytics let's put that knowledge into practice!