After posting on the discussion board and emailing a few classmates, I was unable to get a team. I am perfectly fine doing the project on my own, but if it is a requirement I will need the professor to place me with a team mate.
Team Members: Jacob Walsh, Jacob Damiani, Eric Gagnon
The data: Data was collected in the first half of the 2014-2015 NBA season on shots taken by every NBA player. The response of interest is whether a shot is made or missed. I want to look at how I could use the data available to present to a coach or player that might be interested in the analysis. How can shot selection have an impact on whether the shot is a make or miss? The available predictors of interest are: shot number, period, game clock , shot clock, dribbles, touch time, shot distance, points type, distance to defender, and a few others. The data was found through my Kaggle account: https://www.kaggle.com/dansbecker/nba-shot-logs.
Methods: Binary logistic regression models with mixed predictors will be investigated determining what types of predictors are important for predicting whether or not a shot will be a make or miss. Some continuous predictors such as shot distance will need to be separated into categories, the distribution is clearly bimodal.
Examples of relationships I would like to investigate: Does the shot type (2 or 3 points) effect success? How does the number of dribbles a player takes impact whether or not the shot is made? Does the distance to a defender and shot clock have an impact on shots made? Does the length of time a player holds the ball before shooting impact made shots? Does the game period have any impact? I will be looking at odds ratios and interactions and effects of the different combinations of predictors within the data set to see what I can learn about shot selection.
The presentation is viewable by others.