This project was created as the final project for the Le Wagon Data Science & AI bootcamp in Lisbon

Game of Thrones Survival ⚔️

I ♥️ Game of Thrones. And I ♥️ speculating about how long I would survive in a movie or TV show. So finally we were able to make a website that will create a character based on the user's character traits and then predict how long they would survive in Game of Thrones.

Approach

📊 We used datasets available on kaggle, and did our own webscraping to enhance those datasets with more information.

🧹 The data cleaning was not as easy as expected, as we had to deal with many missing values and duplicates. Our scraped data helped us here, to fill in some of the missing values and give us enough information to impute the rest.

❓ We then moved on to building the quiz to create a character for the user. We created our own dataset that we used to map a user to one of 20 groups or houses Game of Thrones. This was not too complicated, but we needed to make sure that each group had a distinct set of feature combinations, as there couldn't be any duplicates in this dataset.

💀 Once the character creation was done, we built two models to predict whether the newly created character would die, and if so, in which season. Linear models turned out not to work very well, so we settled on using XGBoost.

📜 Additionally, our project involved the development of a storyline and images using generative AI from OpenAI. If the character is predicted to die in a certain season, this will create a story of death and a matching image. And if the character is lucky enough to survive, it will be a short tale of their life.