My attempt at Home-Credit-Default-Risk

This competition was the first time I tried using manual feature engineering in order to boost my prediction score. Although I have had relatively few experiences in data science, I always thought that model creation and machine learning was the most important step. However, after seeing the significant improvement from just a simple feature extraction from additional files; I realized feature extraction is much more time-efficient than perfecting a machine learning model. I followed the guide Introduction to Manual Feature Engineering and implemented this framework with different files in order to feature extract new information. While I didn't do the greatest (4595 out of 7198) I'm happy that I learned something new and even obtained a unique score on the leaderboard!

4595 out of 7198 (My first unique score)