👥 Cluster Analysis Tool
https://tinyurl.com/clusterer
Goal is to find groups (or clusters) of people such that the behaviors within each group are similar to each other, while behaviors across different groups are very different.
For example, In a Streaming Service, like Netflix / Youtube for each person in the platform, we could look at : Minutes watched per day, Number shows/clip in a day, avg length of content, type: drama vs educational vs political, etc..
And the analysis might reveal unique clusters like: Binge tv drama watcher, the once-a-week movie watcher, or the tik-tok watcher, etc..
These translate into specific personas that use the streaming service differently. So once the streaming learns this, they can then customize the experience to better serve these personas.
You might find personas you never thought of that maybe the streaming service is not supporting adequately yet.
This is an unsupervised data science technique, where there's no right answers to start with, the algorithm is not learning from examples, is actually calculating distances between behaviors and figuring out those cluster by itself