Metacritic.com is a review site for movies, videogames, music and tv shows. This project is focused on videogames and is divided into 4 sections:
The first step is to get the data from metacritic website. Game information and reviews are scraped for each title of PlayStation 4, Xbox One and Switch platforms.
Once raw data is available cleaning tasks are done (e.g. dealing with missing values and incorrect formats).
A variety of analysis and visualizations are performed trying to answer different questions about the dataset in order to discover its main characteristics.
Several experiments are run to create a sentiment analysis classifier based on reviews. The objective is to train an algorithm that given a new review can classify it in positive or negative.
https://medium.com/@martinpella/naive-bayes-for-sentiment-analysis-49b37db18bf8
If GitHub have problems rendering a Jupyter Notebook ("Something went wrong. Reload?"), paste the link here: https://nbviewer.jupyter.org/