This repo contains codes for the visualization app in my work: hood2vec: Identifying Similar Urban Areas Using Mobility Networks. My work won Foursquare Future Cities Data Challenge.

This repo has also been pushed to heroku, which hosts the visualization app. This app visualizes each Chicago zip code area' neighbors measured in hood2vec latent space (defined in my paper).

Description of important files:

• index.py: the main file to process the interactive visualization app
• chi_zipcode_to_dist_midday: A pickle file which saves a dictionary for Chicago in midday period. The key is zip code. Value is a list of zip codes; the list is sorted (ascending) by the distance to the zip code in the key; the distance is measured in hood2vec latent space. The method to generate this file is in my paper.