Inspiration: The world of sports is constantly changing. With increasingly high bets being placed to enhance the experience of rooting for our inspirations and models, the method to tackle this wager is often arbitrary. However, with growing emphasis and intensity on each game and tournament, the way we should approach our predictions should be rooted in history, statistics, and data. This simple interaction by people linked by wagers and bets is a thriving business in Las Vegas, Korea, and Liverpool. - it is a universal feeling to want to be engrossed in the sport you are viewing, to take a slice of that stake's cake and to share the same intensity, anxiety, and ultimately triumph as the athletes. Therefore, we wish to expand this model beyond picking lucky numbers and give the audience; and hopefully, the sport's industry, a sustainable business model for individuals and businesses who truly care to live in an athlete's glory.
What it does: Data was collected from the last 4 years, relating to unique key characteristics about athletes and the team they represent, and this data was processed and cleaned. After cleaning the data, data analysis was carried out multi-variable regression models with the help of which it was possible to determine the probability of winning a certain team in a game. The web page was used as a presentation of how this data was obtained and what results were achieved.
How we built it: The data we collected was on in-game statistics from the NBA regular season. A model was built in the language, R, with the total points scored in a game as the response variable and other in-game statistics as the predictor variables.
Challenges we ran into: We had some issues determining the significance of predictor variables, however we were able to use the Akaike Information Criterion to subset the predictor variables to only significant variables. This allowed us to have R-squared and adjusted R-squared values very close to 1. This ensures a quantitative analysis that is absent of bias and unnecessary variability.
Accomplishments that we're proud of: We are proud of the fact that we were able to create a model that allows us to predict the outcome of a NBA game and use that to simulate the NBA Playoffs. We are satisfied to have created a model rooted in present-day data and accurately tuned to include significant variables that have strong correlation to predict the future points and outcome of a NBA game. Additionally, we enjoyed learning, utilizing, and improving upon programming languages that we were briefly familiarized in, combining that with theoretical and practical knowledge from all our backgrounds and interests: sports, statistics, data science, business, analysis.
What we learned: We were able to learn how to wrangle and clean data, create a significant model, and use it to simulate future scenarios.
In the future, we hope to expand and tune this model to become an industry standard. We are determined to add an additional layer to the sports industry and create an experience to closely connect the bridge between athletes to fans and from fans to fans. This statistical model is deeply rooted in data, but its main purpose and core function is human. It's an emotional experience to support, to cry, and to commit to wagering on your idol. This extra dimension to sports is profitable and engaging. We love sports, and all wish to find ways to feel the same vigor oozing through our TV screens and believe this model is one step closer to accomplishing that.
Log in or sign up for Devpost to join the conversation.