The Scenario
You are a commercial real estate broker. You have just listed thirty properties in a Google Sheet — column A has the addresses. A prospective investor called and asked how walkable each property is. Specifically: how many coffee shops, restaurants, and transit stops are within 500 meters of each one. She wants numbers by the end of the call debrief tomorrow morning.
You have used Mapbox before for routing. The POI search API is a different endpoint you have never touched.
The bad version:
- Read the Mapbox Search API documentation, figure out the category query syntax for coffee shops versus restaurants versus transit, and realize each is a separate API call per address per category.
- That is 30 addresses times 3 categories equals 90 API calls, each of which returns a paginated list you have to count.
- Export the results to a CSV, bring them back into the sheet, count by address and category, and paste the counts into three separate columns.
The debrief is at 9 AM. It is 6 PM.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Google Sheet. It reads the property addresses, calls Mapbox category search for each POI type at each location, counts the results, and writes the counts — and nearby names — back into the right columns without any manual API work.
For each property address in column A, geocode it using Mapbox and count how many coffee shops are within 500 meters — write the count into column B and the names of the nearest 3 into column C. Do the same for restaurants in columns D and E, and transit stops in columns F and G.
What You Get
- Column B with the count of coffee shops within 500 meters for each property.
- Column C with the names of the three nearest coffee shops.
- Column D with the restaurant count.
- Column E with the three nearest restaurant names.
- Column F with the transit stop count.
- Column G with the nearest three transit stop names.
What If the Data Is Not Quite Ready
The addresses are missing unit numbers and the geocoder might return building centroids
Unit numbers do not affect the POI search radius — they matter for geocoding precision, not for what is nearby. But you want to confirm the geocode landed in the right place.
Geocode each address in column A using Mapbox. For any address where the geocode match type is not "address" level — write "APPROX" into column H. Then run the 500-meter POI search for coffee shops, restaurants, and transit stops regardless, writing counts into columns B, D, and F and top-3 names into columns C, E, and G.
The investor also wants grocery stores and pharmacies
She expanded the request on the call. Two more categories.
For each address in column A, use Mapbox category search to find counts within 500 meters for: coffee shops (column B), restaurants (column D), transit stops (column F), grocery stores (column H), and pharmacies (column J). Put the nearest 3 names for each category in the adjacent column (C, E, G, I, K).
You want a walkability score that weights the categories differently
You want a single composite metric: transit stops count double, coffee shops and restaurants count once each.
Run a Mapbox POI search for coffee shops, restaurants, and transit stops within 500 meters of each address in column A. Write raw counts into columns B, C, and D. Then compute a weighted walkability score in column E using the formula: (transit count times 2) plus coffee shop count plus restaurant count. Sort the sheet by column E descending.
Full pipeline: geocode, multi-category POI search, score, and highlight top properties for the investor
Geocode all addresses in column A using Mapbox. Count coffee shops, restaurants, transit stops, grocery stores, and pharmacies within 500 meters of each. Write counts into columns B through F and name lists into columns G through K. Compute a weighted score in column L using transit count times 2 plus the sum of all other categories. Sort by column L descending. Mark the top 10 properties with "INVESTOR PICK" in column M.
One prompt handles geocoding, all five category searches, scoring, sorting, and highlighting.
Try It
Get the 7-day free trial of SheetXAI and open your next property analysis sheet — paste in the addresses and describe the POI categories you care about, and SheetXAI will call Mapbox and write back counts and names for each. The isochrone article covers catchment area analysis if you need reachable-zone polygons instead of point counts. The Mapbox hub has the complete list of geographic tasks.