Inspiration
The film industry, as with most entertainment industries, can be very unpredictable. As a film producer, it can be difficult to know which movies will be a hit among viewers and which ones will be box office flops. For this reason, it can also be challenging to know how much a film is really worth investing in. FilmFam aims to simplify this decision.
What it does
The website takes in four parameters with respect to a movie that the user/producer wishes to make, namely its genre, desired revenue, desired score (imdb rating) and duration. Through the implementation of a machine learning algorithm, it predicts the best budget the producer should choose for the film in order to achieve the desired revenue.
How we built it
We first spent time crawling and extracting data from APIs and websites. Then we tried many different Machine Learning models and tried to increase the accuracy of those models. After choosing the models, we wrote frontend and server code.
Challenges we ran into
The main issues we faced were maximizing our accuracy without overfitting the training data and finding data to train and test our model on. Additionally, we faced some difficulty in connecting the different components together (the trained model, server and frontend).
Accomplishments that we're proud of
We're glad we got everything to work in the end. We are particularly happy that our model worked well, we were able to extract/crawl more training/testing data than we expected to find and we were able to successfully connect the ML model and server to the frontend.
What we learned
Given that we had three first time hackers, a lot was learned throughout the event. Each one of us learned about various different components of Machine Learning, server-side scripting and web development.
What's next for FilmFam
We hope to improve the accuracy of our model by trying new techniques such as convolutional neural networks and recurrent neural networks. Further, we hope to be able to predict more parameters that may seem to have a correlation with the movie data available (such as the release date or location or cast to maximize profits).
Log in or sign up for Devpost to join the conversation.