Getting insights from the data

1. Getting insights from the data

Welcome back! We'll continue our analysis to find even more insights in the data.

2. Cluster analysis

One of the most common descriptive and exploratory techniques is cluster analysis. Cluster analysis aims to find naturally occurring groups in the data, called clusters. Possible cluster analysis applications you might have encountered are customer segmentation, making a classification, or identifying subgroups. Regardless of the application, there are typically two main steps in cluster analysis: finding the optimal number of groups and investigating the characteristics of those groups.

3. Finding the optimal number of groups

With this type of analysis, we often don't know the optimal number of groups that are present in the data. So how do we find the optimal number of groups in the data? There are several tools we can use for this. The data scientist in your team can calculate specific statistics and make plots to help find distinct groups. Often it is also useful to use domain knowledge; the HR expert in your team can help evaluate whether the groups found by the data scientist make practical sense. In reality, it can be challenging to find an exact solution. The groups might not be that distinct, for example, or there might be no groups at all, or depending on the case, it might make sense to make a further division into subgroups. For this reason, the final solution in cluster analysis is often the result of teamwork between the data team and the business.

4. Characteristics of the groups

Once we have an idea of how many groups there are, we can move on to the next step: investigating the characteristics of each group. In this step, we are primarily interested in what separates the different groups. For example, take a look at the picture; we can see that the second group is not happy with the current remote working policy, while the first group rates the police higher, and the third group is somewhere in between. If there are more than two groups, we can also look at common characteristics between specific groups. We can also check out some variables of interest of questions we had beforehand. For example, older employees have different expectations from remote working than younger employees and can, therefore, generally be found in different groups.

5. Let's practice!

Over to you now! In the following exercises you'll investigate whether there are separate groups of employees that experience remote working differently.

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.