-
-
Generated map showing 2 recommendations for new food amenity locations in St. Louis
-
Map generation options
Inspiration
When I heard the theme "Smart Cities", nothing immediately came to mind. The term makes me think of something futuristic (and unrealistic), not anything that I particularly want to make. But when I took a few more seconds and considered what "smart city" actually means, I realized that anything that can improve the environment people live in contributes to a smart city. I like food, and also need it to survive, so having food around me seems smart. I remember at some point hearing about food deserts, and some census data pertaining to it, and decided to use that as the basis of my hack.
What it does
I made an app to identify optimal locations for a new food source. The user can input the type of food source (either amenity like a restaurant or cafe, or a shop like a supermarket or convenience store), as well as the specific cuisine if they entered amenity. The program will then find a selection of optimal locations (within Missouri by default, since that's the census data I used) for the user to set up their new food source. This selection is calculated by first finding the regions that are considered to have low food access and are far from food sources matching the user's, and then for the chosen regions finding a specific building close to populated areas and far from other food sources. The chosen buildings are then marked on a visual map displayed to the user.
How I built it
I made it in python using geopandas. and I lot of confusing and asking chatGPT how python works.
Challenges I ran into
The Overpass API was a bit inconsistent to work with, which forced me to implement handling for API failures (which maybe I should have implemented regardless but sometimes I just want to be lazy yn). This was also my first hackathon project, and it was a bit of a challenge getting started and figuring out what I wanted to make.
Accomplishments that I'm proud of
I'm mostly just proud that I did this hackathon, and I made a semi-complete product.
What I learned
I still don't like python. Map data is pretty interesting (because sadly the earth is not flat (and also just because maps are interesting)).
What's next for Food Location Suggester
I was hoping to implement some sort of LLM to supplement OSM data by generating a cuisine tag when it was missing based on name & web info. The biggest improvement that is needed to make this more than just a demo is to be able to easily select a region for the search. Currently it is using pre-downloaded data specific to MO, and doesn't have any options to reduce the scope either.
Log in or sign up for Devpost to join the conversation.