CreatingTheNext

Inspiration

Currently, the government's resources are limited, and too much is being spent on unemployment. Politicians make mistakes in allocating too much or too little amount of money for unemployment, especially with the time constraints that they are subject to. We wanted to assist them by creating a tool that analyzes past unemployments rates and other economic factors to produce an appropriate amount of spending the government should use.

What it does

Our website uses the power of data science to first analyze unemployment rates and homelessness rates, as well as other macroeconomic factors affect them (including GDP, CPI). Based on those data, the website employs a machine learning model to generate a formula for calculating the appropriate amount of the GDP the government should reserve for unemployment spending.

How we built it

We first processed our data from multiple datasets provided by hacklytics.io, as well as some datasets from our own search (including FRED).
After storing this data into MongoDB, we processed the data using pandas and NumPy.
We generated visualizations for the data using seaborn, a data visualization package that is built off of the popular package matplotlib.
We created a machine learning model, specifically a multiple regression model, using scikit-learn. The input is based on the current unemployment rate, homelessness, and other macroeconomic factors (including GDP, CPI), and outputs an amount that the government should spend on homelessness.
Finally, we created a website using Flask and React to deploy and display our results.

Challenges we ran into

Because we went through various data science technologies, we not only had to learn how to use them, but also combine them to form the meaningful product we wanted to create.

Accomplishments we're proud of

Using visualization tools to display our results, learning about statistical methods, and collaborating with each other to accomplish more!

What we learned

Since we were beginners in data science, we learned a lot about the fundamentals of data science, including analysis techniques and visualization tools. We all learned about the seaborn visualization tool, as well as methods in scikit-learn to analyze our data.

What's next for CreatingTheNext

Using historical data to analyze the future is quite powerful, but we would like to also incorporate sentiment analysis for CreatingTheNext given more time to analyze the current sentiment on unemployment and homelessness appropriations.