Determines important characteristics of highly rated video games, and uses nltk and sklearn for sentiment analysis and genre classifcation through Machine Learning models.
As someone who loves video games, in this project, I wanted to explore what were characterstics of video games that led to higher rankings. After all, what makes a video game "good" is incredibly subjective. I viewed change in trends over 40+ years of time across the video game industry, looked at coorelation matrices between many variables including a sentiment analysis score based on the reviews that people left for each game, and lastly created a genre classification model after testing the accuracy of 4 different machine learning models known to perform well with classification based on the numerical features I engineered.
Overall, this Analysis involves NLP Techniques, Statistical modeling of Video Game Trends, Coorelation Matrices, Sentiment Analysis, various data visualizations, Time Series Data Analysis, and lastly utilizing Machine Learning Models and ensembling techniques to create a genre classification model.
Github doesn't properly load my Jupyter notebook due to some of the libraries I worked with, so to view my entire analysis, please use the following link: https://nbviewer.org/github/AdishSundar/VideoGame-Analysis-Project/blob/main/VideoGameAnalysis.ipynb