The aim of this project is to see if there is any correlation between sales and aggregate review scores for video games and how they focus their development. The data used for this project is based off harvested elements from the website Metacritic (data collected in 2016), and has gathered details on the names of games, their year of release, regional and global sales figures, aggregate critic review scores, and aggregate user scores. The data was then cleaned to remove any games that only had partial data present about them as they would not provide any meaningful statistic to the study, the criteria for this removal was missing sales data, review data or year of release, as all of these are major factors in understanding we are seeing certain patterns. After cleaning the data down to approximately 7000 entries from the original 16720 entries, it was tagged manually between three categories, first looking at a 50/50 split between story and gameplay focus; meaning that certain games could be both gameplay focused, and story focused and providing an additional sorting category for whether the game is part of a series. For the sake of this study, a series will be understood as a collection of games that follow the same character throughout their story, exist within the same world and have relevance between themselves, or feature character development elements between them. This manual tagging and discarding of entries has created some bias within the data, as I am personally not familiar with all of the games present on the list, and this has resulted in some game genres being tagged as gameplay focused simply because I don’t know much else about it. With more time and research, I would be able to eliminate this further, but it simply was not possible to fix all of the inconsistencies on my own. Another issue with the tabular data used is that due to games being released on multiple platforms, they can appear in the visualisations multiple times. Along with this, older games have relatively low sales numbers due to an ever-increasing number of players worldwide. To accommodate for this for this bias, sales will not be filtered to only feature games that have sold more than a specific number of units. All of the visualizations of the data were generated with Voyager 2 (vega.github.io/voyager) as it was the most robust tool that allowed me to change between different views in order. The data itself was cleaned with OpenRefine v3.1. The filters used are available in the file titled “Voyager2FilterExport.txt”

This project was made for a Master's level university course.