Coffeeshop Site Selection Using Geospatial Analysis

Yelp API Data Collection Script

Complete Python implementation for scraping competitor location data:

import requests
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from YelpAPIKey import get_key
from copy import deepcopy

# Configure pandas display
pd.options.display.max_columns = 2000

# Define API Key, Endpoint, and Headers
API_KEY = get_key()
ENDPOINT = 'https://api.yelp.com/v3/businesses/search'
HEADERS = {'Authorization': f'bearer {API_KEY}'}

# Define search parameters
PARAMETERS = {
    'term': 'coffee',
    'limit': 50,
    'offset': 400,
    'location': 'Saint Louis'
}

# Collect all businesses with pagination
businesses = []
for offset in np.arange(0, 1000, 50):
    PARAMS = {
        'term': 'coffee',
        'limit': 50,
        'offset': offset,
        'location': 'Saint Louis'
    }
    response = requests.get(url=ENDPOINT, params=PARAMS, headers=HEADERS)
    data = response.json()
    
    for item in data.get('businesses'):
        businesses.append(item)

# Process categories (maximum 3 per business)
for business in businesses:
    if len(business.get('categories')) == 3:
        business["cat_1"] = business.get('categories')[0].get('alias')
        business["cat_2"] = business.get('categories')[1].get('alias')
        business["cat_3"] = business.get('categories')[2].get('alias')
    elif len(business.get('categories')) == 2:
        business["cat_1"] = business.get('categories')[0].get('alias')
        business["cat_2"] = business.get('categories')[1].get('alias')
        business["cat_3"] = 'None'
    else:  # length == 1
        business["cat_1"] = business.get('categories')[0].get('alias')
        business["cat_2"] = 'None'
        business["cat_3"] = 'None'

# Extract location and coordinate data
for business in businesses:
    business["street_address"] = business.get("location").get("address1")
    business["zip_code"] = business.get("location").get("zip_code")
    business['latitude'] = business.get('coordinates').get('latitude')
    business['longitude'] = business.get('coordinates').get('longitude')

# Convert to DataFrame and clean data
df = pd.DataFrame.from_dict(businesses)

# Remove unnecessary columns
drop_columns = [
    'id', 'alias', 'image_url', 'url', 'categories', 
    'coordinates', 'transactions', 'location', 
    'phone', 'display_phone', 'is_closed'
]
df.drop(labels=drop_columns, axis=1, inplace=True)

# Filter for coffee-related businesses
df_coffee = df.loc[
    (df["cat_1"].str.match('coffee')) |
    (df["cat_2"].str.match('coffee')) |
    (df["cat_3"].str.match('coffee'))
]

# Remove chain competitors (focus on independent shops)
df_coffee = df_coffee[df_coffee['name'] != "McDonald's"]
df_coffee = df_coffee[df_coffee['name'] != "7-Eleven"]
df_coffee = df_coffee[df_coffee['name'] != "Dunkin'"]

# Create cost categories from price indicators
df_coffee["cost"] = np.where(df_coffee["price"] == "$", 1,
                   np.where(df_coffee["price"] == "$$", 2,
                   np.where(df_coffee["price"] == "$$$", 3, np.nan)))

# Export cleaned data for ArcGIS Pro import
df_coffee.to_csv("yelp_stl_coffee_data.csv", index=False)

print(f"Data collection complete: {len(df_coffee)} independent coffeeshops identified")
print(f"Average rating: {df_coffee['rating'].mean():.2f}")
print(f"Price distribution: {df_coffee['cost'].value_counts().to_dict()}")

Data Processing Summary

The script processes Yelp API responses to:

• Handle pagination to collect comprehensive competitor data
• Parse business categories and extract coffee-related establishments
• Clean and standardize location data for GIS import
• Filter out chain competitors to focus on independent coffeeshops
• Create cost classifications for market analysis