Data Here!

http://mvp-prediction.appspot.com/

Inspiration

Passion for Basketball and a strong desire to make a machine learning project. We are a team of 4 individuals who have never done anything machine learning related until this project. We saw this as an opportunity to grind for many hours in order to learn something new together in an environment that we like.

What it does

Our main functionality is to predict future NBA MVPs based on previous year data that we scraped up which was used to train our models,

How I built it

We primarily used python and some libraries / APIs to create Linear Regression, Random Forest Regression, and Gradient Boosting regression models. We had to first scrape data from internet which dated back to the 1970-1980s. Using this data we created a Correlation Matrix to determine how much of this information was relevant and accurate to train our models. ex: true shooting. After our model was trained, we then saw the accuracy of the predictions it could make and the actual predictions that it made. We tested it on the 2020 season and got results that WOULD make sense. We also tested a past years MVP selection pool and got accurate results.

Challenges I ran into

The biggest Challenge we faced was that this was everyone's first Machine Learning Project. We came to the Hackathon with an idea but not too much of a clue on how to make it. Throughout the time we spent here, we learned as we made the project. One of our biggest turning points was going to the Machine Learning 101 workshop. We learned so much from this workshop and the speaker was great. We can confidently say that this workshop paved the way for us to finish the project on time. We learned which models we should create (in general). Eventually we hit more roadblocks and confusion until one of the Mentors from Softeon gave us more insight on which models were more appropriate to use for our project. From there we made insane progress and eventually were able to finish this project. Our last and final challenge was trying to register a domain for our website, using the codes that we received from MLH. However the CEWIT network was blocked from registering domains for whatever reason and until now, we still haven't been able to register a domain. We had a long conversation with customer support but nothing resolved our issue.

Accomplishments that I'm proud of

We are all proud on how much we learned. After developing these machine learning models and getting results that are accurate, we are all now very eager to expand our knowledge within the field.

What I learned

We learned how to use different APIs and libraries in Python to make machine learning models. We also learned which machine learning models were appropriate for a given situation. Finally we learned one of the most important skills that a programmer could have and that would be strong collaboration skills and teamwork.

What's next for Prodict

Next up, we want to be able to implement a way to retrieve information directly from source websites in order to update our ranking predictions dynamically.