With the 2020 U.S census looming over the horizon, I thought it would be appropriate to give the past data kiss goodbye. On a serious note, I just thought it would be interesting to see what patterns exist between poverty, social services, and other factors.
What it does
Ideally U.S Census Poverty Analysis will display data about poverty for the County selected. The application may use some kind of machine learning or linear regression to predict future poverty data. The app will also find outliers where the rates of poverty vary greatly over time. I hope that policymakers can learn from outliers to make better decisions to end suffering.
How I built it
We got data from the Census website, and cleaned it. Then we used streamlit to present the data as line charts for counties or states selected by the user.
Challenges I ran into
It was our team's first time using streamlit, and we had some difficulty using it. I had trouble with the datasets, as some functions to delete columns would not work. We didn't get far enough to implement linear regression and machine learning.
Accomplishments that I'm proud of
Before this project, I knew some of the syntax of the python, but not much more. I wrote my first python scripts for this project, and started to learn tools I've never used before.
What I learned
streamlit, matplotlib, teamwork, python shell. I honestly learned a lot.
What's next for U.S Census Poverty Project
We need to improve performance for retrieving the data, before we can go a lot further. Moreover, I am interested in learning a tool called "CSPro". Once we/I get more familiar with how to query census data, we could maybe apply machine learning or linear regression to make predictions about poverty.