This was a project for the TAMU Datathon 2020.
Caroline Kostrzewa, Lori Kolaczkowski, Veronique Marcotte
What it does
It clusters cities based on user input. It shows where that user's ideal location would be in relation to other cities using information about what they think is most important.
Challenges we ran into
Getting good data was a challenge. We also had some long discussion on the best approach -- should we use PCA to reduce the number of feature dimensions? Should we use clustering to group similar cities? How should we apply the weights (i.e. importance of each factor to the user)?
What's next for City Search Tool (TD 2020)
Our big plan for the future is to make this into an interactive web app (via R shiny). However, we also want to add more factors and look more deeply in the weighting of each factor.