Context As the season has come to an end and at the moment we are already deep in playoff basketball, I wanted to take a look and see if I can at any way get to some data so I can predict the MVP of 2018-19 season. After a quick search, I came across all mvp votings since 1968-69 up to this past seasons on [basketball-reference](https://www.basketball-reference.com). I wrote a scraper and got the data. I also got the data for current season. However, I scraped only the data from 1980-81 season up to now because that's when the media started to choose MVP of the league. Content The `mvp_votings.csv` represents the train data. It holds various basketball statistics. You can view some of the descriptions of the stats in my [medium post](https://towardsdatascience.com/predicting-2018-19-nbas-most-valuable-player-using-machine-learning-512e577032e3) The target value for regression can be `award_share` column which represents the share of the votes that the players have won. Acknowledgements All of the data is owned by basketball reference, and I do not own any of the data. Image belongs to [nba.com](https://www.nba.com/video/2019/03/25/james-harden-giannis-antetokounmpos-top-plays-season) Inspiration What is the most important statistic which defines how will be the MVP? What are your predictions for this season? How did the most important feature change over the year? How big of an impact does a team's win percentage hold with all other features?