Find Your Degree is an advertising-supported site. Featured or trusted partner programs and all school search, finder, or match results are for schools that compensate us. This compensation does not influence our school rankings, resource guides, or other editorially-independent information published on this site.

What is Cluster Analysis?

Cluster analysis is a concept that is often found in statistics courses, and that is present in the daily practice of many fields, including medicine and social science. While cluster analysis can seem like a confusing topic, it is really a basic organizational technique that helps scientists and analysts understand how things may be related to each other. A basic understanding of the underpinnings of this statistical tool make it less intimidating for students delving into the many fields that require research or data analysis.

Definition of Cluster Analysis

In it’s simplest form, cluster analysis is a method for making sense of data by organizing pieces of information into groups, called clusters. Data points can be survey responses, images, living organisms, chemical compounds, identity categories, or any other observable type of data that helps professionals explore problems and questions. Clusters can be made up of any number of data points that are related in any number of ways that are defined by the researcher. Algorithms are often a helpful tool for determining the data points that belong within a cluster. Most analysts will select the clustering model that best fits their data and choose an appropriate algorithm based on the model.

Clustering Models

Cluster analysis is a broad umbrella for many different methods of statistical analysis that create “clusters” through different organizational means. While there is no set definition for what comprises a cluster, there are several common models for assembling various types of clusters. The selected model will vary depending on the needs of the researcher or the more general tenants of their field of study.

Some of the most common cluster models are hierarchical, density, and distribution. Hierarchical clustering uses an algorithm that connects data points by distance. The idea behind the hierarchical model is that data points nearer to one another are more related than ones farther away. For each set of points, analysts must determine the desired amount of distance required for points to be contained within a single cluster. Density clusters are defined by dense points within the field of data. The sparser areas separating clusters are not grouped within this model. Distribution clustering is most closely related to statistics, and mandates that clusters are determined via the distribution origin of each data point. Points belonging to the same distribution will be grouped together.

Cluster Analysis Professions

Biology, social science and marketing are just a few examples of professional fields that employ cluster analysis. Within biology, for instance, scientists use cluster analysis to group plant organisms within a genus or family that display similar attributes. Social scientists use cluster analysis to determine areas where certain types of crime occur at a higher rate. Population analyses and educational studies are other areas of social science that involve cluster analysis. Columbia University emphasizes the importance of cluster analysis to marketing and business professionals seeking to identify target groups for particular products and services.

Researchers and analysts require tools to find meaningful results from sets of data. Because of its ability to be manipulated for a numerous variety of research purposes, cluster analysis is a valuable resource for any professional working with data points.

Related resource:

20 Best Data Science Bachelor’s Degree Programs