Finding locations to open a gym in Brooklyn, NY

This project uses geospatial analysis + clustering to recommend potential new gym locations in Brooklyn that are close to the Brooklyn city center and far from existing gyms/fitness centers.

Full report is included in this repo: Report-Finding-locations-to-open-a-gym-in-Brooklyn.pdf

Goal (Constraints)

• Search area: within 5 km of Brooklyn city center
• Competition constraint: candidate locations must be ≥ 2.5 km away from an existing Gym/Fitness Center (and similar venues like boxing clubs/gym pools)

What it does

• Loads Brooklyn neighborhood coordinates (NYU GeoJSON dataset)
• Uses Foursquare Places API to find existing gym/fitness venues
• Generates a dense grid of candidate points (~200 m spacing) around the city center
• Filters candidates that satisfy the distance constraints
• Applies K-Means clustering to group candidates into “zones”
• Outputs 10 zone centers as recommended location anchors (optionally reverse-geocoded to addresses)

Tech Stack

Python, pandas, numpy, scikit-learn, folium, requests
(Optional) pyproj/geopy for coordinate + distance utilities

Repo contents

• Report-Finding-locations-to-open-a-gym-in-Brooklyn.pdf – full write-up
• notebooks/ – analysis notebook (if included)
• src/ – helper modules/scripts (if included)
• outputs/ – maps / tables (if included)

Running (typical)

1. Install dependencies

   pip install -r requirements.txt

2. Add API keys (example)

.env

• FOURSQUARE_API_KEY=your_key_here
• GOOGLE_MAPS_API_KEY=your_key_here # only if reverse geocoding is used

3. Run the notebook