🗺️ Google Maps Scraper

Dependencies

pip install playwright
python -m playwright install chromium

sudo apt install -y libnss3 libatk-bridge2.0-0 libdrm2 libxkbcommon0 libgbm1 libasound2

Cara Pakai

CLI (direct)

python gmaps_scraper.py "coffee shop" "Yogyakarta" --max 20
python gmaps_scraper.py "restoran" "Jakarta Selatan" --max 10 -o hasil.csv
python gmaps_scraper.py "tempat wisata" "Bali" --max 15

Via AI Agent

"Cari coffee shop di Jogja dari Google Maps"
"Scrape restoran di Jakarta Selatan rating >4.5"
"Tempat wisata di Bali yang buka 24 jam"

Arguments

query

location

--max

--output, -o

--headless

--no-headless

--delay

Output Format

Teknis

1. Buka Google Maps search URL
2. Scroll results panel untuk load lebih banyak listing
3. Klik setiap card → extract data dari detail panel
4. Parse: Name, Rating, Reviews, Price, Category, Address, Phone, Website, Hours
5. Save ke CSV + JSON

Limitasi

• Tanpa login Google: max ~20 results per search
• Rate limit: jangan scrape terlalu cepat (default delay 1.5s)
• Data bisa berubah sewaktu-waktu (Google update struktur)
• CSS selectors mungkin perlu update kalau Google Maps redesign

Troubleshooting

playwright not found

pip install playwright && python -m playwright install chromium

chromium failed to launch

sudo apt install libnss3 libatk-bridge2.0-0 libdrm2 libxkbcommon0 libgbm1 libasound2

timeout

no results

0 data / CSS selector error

button

a

redirect ke detail page

Changelog

2026-04-16 — Robust Selector + Redirect Fix

• Fix: Selector

  a[class*="hfpxzc"]

  → auto-detect

  button

  (new) /

  a

  (old) / aria-label fallback
• Fix: Handle redirect ke detail page → auto-retry dengan coordinate URL + search box
• New:

  get_listing_cards()

  — unified card detection
• New:

  is_detail_page()

  — detect if redirected to single listing
• New: Debug screenshot saved if no cards found

2026-04-16 — URL-Based Extraction (v2)

• Fix: Click-based extraction only got 7 results (DOM breaks after clicking)
• New:

  get_card_links()

  — collect all place URLs first, then visit one-by-one
• New:

  extract_from_detail()

  — extract data when already on a place URL
• New:

  extract_listings_click()

  — fallback click-based approach
• Result: 7 → 45+ results per search
• Default:

  --max

  changed from 20 to 50