Marvel Characters Clustering

Marvel Characters Clustering

Journal
Year
2022
Co-authors
Tags
Report
Description
I employ unsupervised learning techniques to identify hidden patterns among Marvel characters. Superheroes are representative of a society's desires and aspirations, thus being an interesting topic for social research. In order to shed light on such aspirations and societal ideals, I exploit both Clustering Analysis and Categorical Principal Component Analysis.
Field
MachineLearning
notion image
 
Superheroes have the possibility to embody anything a human could never be. They are thus created as an echo of the desires of a particular society in a given historical moment. The greatest example of this phenomenon is Superman (belonging to the DC Universe): the first comic was released on 10th June 1938, while Europe was distraught by totalitaristic regimes. In that context, he was created to epitomize the idea that strength and power could be used for good, and to give the population the hope that there was someone, despite fictional, that could fight the dictators. The research purpose of this work is thus to identify these concepts and ideas, peculiar to the time and society in which Superheroes are created, by analysing the ”Marvel Characters Data” dataset available on Kaggle.