Sweden GBP Discovery Scraper

Apify actor that systematically discovers all Google Business Profiles in Sweden using an adaptive spatial grid and Playwright browser automation.

How it works

1. Input: An array of grid cells, each with coordinates (lat/lng) and a zoom level
2. For each cell × search term: Opens Google Maps at the specified coordinates, searches with broad Swedish terms
3. Scrolls through all results to load every listing in the viewport
4. Extracts business data: name, address, phone, website, placeId, CID, category, rating, review count
5. Deduplicates across all searches by Google Place ID
6. Reports saturation: If any search returns ≥120 results (Google's cap), it flags the cell for subdivision

Input

Field	Type	Default	Description
cells	Array	required	Grid cells: { cellKey, lat, lng, zoom }
searchTerms	Array	8 Swedish terms	Broad search terms for discovery
maxScrolls	Integer	8	Max scrolls per search (each loads ~20 results)
visitDetailPages	Boolean	true	Click into each listing for phone/website/CID
language	String	sv	Google Maps language
maxConcurrency	Integer	3	Parallel browser pages

Example input

{
    "cells": [
        { "cellKey": "stockholm_center", "lat": 59.3293, "lng": 18.0686, "zoom": 14 },
        { "cellKey": "malmo_center", "lat": 55.6049, "lng": 13.0038, "zoom": 14 }
    ],
    "searchTerms": ["företag", "restaurang", "butik"],
    "visitDetailPages": true
}

Output

Each discovered business is pushed to the default dataset:

{
    "placeId": "ChIJx8uLN7h3X0YR...",
    "googleCid": "1234567890",
    "name": "Restaurang Example",
    "address": "Storgatan 1, 111 23 Stockholm",
    "phone": "+46812345678",
    "website": "https://example.se",
    "latitude": 59.3293,
    "longitude": 18.0686,
    "category": "Restaurang",
    "categories": ["Restaurang"],
    "rating": 4.5,
    "reviewCount": 123,
    "googleMapsUrl": "https://www.google.com/maps/place/...",
    "cellKey": "stockholm_center",
    "searchTerm": "restaurang",
    "resultIndex": 0
}

Search metadata (per cell × term) is stored in the key-value store under SEARCH_META.

Adaptive grid approach

Sweden is divided into ~1,100 quadtree cells where each cell contains ≤80 known businesses. Dense urban areas (Stockholm, Göteborg, Malmö) get small cells at zoom 14; sparse rural areas get large cells at zoom 7-10.

If a search returns ≥120 results (Google Maps' maximum), the orchestrator subdivides that cell into 4 smaller cells and re-queues them.

Coverage: 1,103 cells × 8 terms = 8,824 searches → estimated 300K-500K unique GBPs discovered.

Tech stack

• Apify SDK + Crawlee — actor lifecycle, dataset storage, request management
• Playwright — headless Chrome for Google Maps rendering and interaction
• TypeScript — type-safe extraction logic

Development

cd sweden-gbp-scraper
npm install
npm run dev    # Run locally with Apify storage emulation

Deployment

apify login
apify push