Project Details

What do we do as Data Scientists? We look at the data, do some cleaning and feature engineering, and run some predictive modeling to train a model to create some kind of output. The goal of this model is to predict whether or not a Kickstarter campaign will be successful. We looked at what features were the most important and isolated those and ran through a RandomSearch to produce our current working model.

Originally we started out with a 5 layer neural network, but constantly running models proved to us that the simpler the better in this case.

Our baseline accuracy was a 61% while our model was easily achieving 75% accuracy or higher.


What We Did

Predictive Modeling
Data Scraping

Analyzing data isn’t easy, let me give you a fresh pair of eyes!