Project for Datahacks 2025
https://video-game-classification.streamlit.app/
Due to the sudden emergence of indie games toppling the video game charts over the past couple of years, we had a rough idea of trying to do a project based on video games. We were inspired after looking at a couple of reviews from looking at one of the biggest video game distributors in the world Steam.
We created a classification model that took in any review that we type in and it would generate a video game genre that would best fit that review description.
We used Kaggle to find two datasets that contained video game reviews of all sorts of games and used pandas to wrangle and clean the data. Then, we picked different genres for our model to classify reviews into and used sklearn to create and train our data.
When we were structuring our genres, there were a lot of concatenation of other genres causing us to have to decide on which genre should take precedence or find a way to work multiple genres into our model. Additionally, choosing the genres that the model would take was another challenge because if we used a genre that was too broad like "Adventure" or "Action" would cause our prediction model to either default to those two genres and nothing else.
Getting hands on experience with streamlit, sklearn, and pandas
How to make a more completed product