Us
We are the Linear Bounded Automabros, we are in Tier 1. The topic we chose to work on was Topic 1: Use the World Bank's Open Sourced Data.
Inspiration
We wanted to create a single page web application that could demonstrate the value of education in society.
What it does
Ginius leverages data from The World Bank Data Catalog, trains two regression machine learning model, predicts future income inequality (GINI index) based upon national educational statistics. This data is then visualized and the user is given the opportunity to explore how different factors can manipulate that prediction.
How we built it
Ginius was built using a React front-end with a Flask back-end and also implemented sci-kit learn as well as Pandas, and NumPy.
Challenges we ran into
The World Bank data we found was quite sparse and required extensive data cleaning in order to be suitable for machine learning, this took considerable time. We encountered some CORS issues that forced us to make some last minute decisions, although all worked out in the end.
Accomplishments that we're proud of
We are extremely proud of our clean UI design and we are quite pleased by the relative accuracy of our prediction model. These two elements came together to create what we think is an impactful submission.
What we learned
Data is not always as clean or organized as you might want, nor is it often in the format you would like it in. We learned real world applications of data cleaning and searching for appropriate data in a large data catalog.
What's next for Genius
Add support for more countries. Along with that, we would clean and prepare more training data for our models to further generalize, allowing them to more effectively predict GINI Indices. Lastly, we also would like at add a dark mode.
Log in or sign up for Devpost to join the conversation.