JP Morgan Chase Data for Good Hackathon

  • Category: Machine Learning/Artificial Intelligence/Data Science
  • Date: October 2023
  • Code: Due to restrictions applied by the hackathon, I cannot share the code

Project Details

5 students and I, answered MoveToProsper's, a company whose goal was moving improvished families into a safer, prosperous neighborhood, question of "What cities in Ohio that we should look to move families out of and into," using data analysis and machine learning techniques. The code we developed identified optimal pairings using a variety of factors (poverty rate, crime rate, average median househould income, etc.) of impoverished and affluent zip codes in Ohio to help MoveToProsper determine new cities to target. It achieves this by clustering zip codes into two groups using KMeans, calculating Haversine distances between the clusters, and scoring the pairs using L2 and Cosine distance metrics. The top 40 pairings with the highest scores are then extracted and saved for further analysis, providing data-driven recommendations for strategic decision-making. From these top 40 pairings, we did human qualitative analysis on census data. From the analysis and optimization function we identified the top 3 cities and their pairings that MoveToProsper should focus on.

The project includes the following key features:

  • Python, Pandas, Scikit
  • Machine Learning
  • Exploratory Data Analysis
  • Leadership

Combining these tools, techniques, and teamwork, our team received the honor to be crowned Fall 2023 DFG's Hackathon winners!