Tony Wang9 min readWhat Google Maps Actually Sees in China: Real Businesses, Not Landmarks
9 of China's top 10 Google Maps categories are real businesses — vs. 4 of 10 in Japan, where shrines, temples, and parks dominate instead.
Pull Japan's largest Google Maps categories and you get a tourism map with businesses mixed in: 6 of the top 10 are shrines, temples, historic sites, parks, apartment complexes, and community centers. Run the identical query scoped to country=China and the picture flips: 9 of the top 10 categories are recognizably, unambiguously commercial — restaurants, shops, salons, pharmacies. Same dataset, same methodology, opposite mix.
What's actually in the top 10
Only the uncategorized bucket (records with no category value at all — 289,459 of them, 3.9% of China's total) breaks the pattern, and it isn't a landmark category, it's missing data. Widen the lens to the top 50 and the commercial skew holds at scale: excluding that uncategorized bucket, only 319,796 of 4,208,439 listings in the top 50 categories — 7.6% — are non-commercial places (government agencies, ATMs, elementary schools, "halls," public toilets, parking lots).
The mirror image of Japan
The same cut run on Japan's data found the opposite: 40.1% of the top-50-category mass there was non-commercial — shrines, temples, historic sites, parks, apartment complexes, bus stops, ATMs, and government offices. Line the two up:
Japan's Google Maps data reads as a tourism-and-civic-landmark map with businesses mixed in — a Shinto shrine and a beauty salon get the same treatment, a pin with hours, a rating, and a review count. China's reads as a straightforward commercial directory. Neither is "wrong" — Google Maps indexes whatever real-world points of interest get mapped and categorized in a given country, and that mix is shaped by the same forces documented in the companion piece on China's near-total absence of reviews and websites: a market where the underlying data collection is asymmetric to how Google Maps ordinarily builds out a place's category and detail signals in markets it fully operates in.
Show all 50 categories
| Category | Listings | Type |
|---|---|---|
| 中餐馆 (Chinese restaurant) | 549,713 | Business |
| (uncategorized) | 289,459 | Unknown |
| 快餐店 (Fast food restaurant) | 164,689 | Business |
| 商店 (Shop) | 156,143 | Business |
| 便利店 (Convenience store) | 155,534 | Business |
| 餐馆 (Restaurant) | 132,844 | Business |
| 美容院 (Beauty salon) | 127,401 | Business |
| 服装店 (Clothing store) | 118,238 | Business |
| 宾馆 (Hotel / guesthouse) | 116,972 | Business |
| 药店 (Pharmacy) | 115,163 | Business |
| 杂货店 (Grocery store) | 108,754 | Business |
| 糕点店 (Bakery) | 97,211 | Business |
| 幼儿园 (Kindergarten) | 95,250 | Business |
| 企业办公场所 (Corporate office) | 92,348 | Business |
| 美发沙龙 (Hair salon) | 91,237 | Business |
| 公司 (Company) | 88,907 | Business |
| 医院 (Hospital) | 87,944 | Business |
| 培训中心 (Training center) | 83,220 | Business |
| 汽车修理和保养 (Auto repair & maintenance) | 81,132 | Business |
| 汽车修理厂 (Auto repair shop) | 80,838 | Business |
| 足部按摩店 (Foot massage shop) | 78,084 | Business |
| 建材商店 (Building materials store) | 76,897 | Business |
| 制造商 (Manufacturer) | 73,751 | Business |
| 家具城 (Furniture mall) | 72,461 | Business |
| 卖场 (Hypermarket) | 68,500 | Business |
| 健身房 (Gym) | 66,985 | Business |
| 房地产中介 (Real estate agency) | 64,832 | Business |
| 鞋店 (Shoe store) | 63,762 | Business |
| 洗衣店 (Laundry) | 61,309 | Business |
| 政府机构 (Government agency) | 60,331 | Place |
| atm (untranslated) | 60,152 | Place |
| spa (untranslated) | 59,349 | Business |
| 蔬菜批发市场 (Vegetable wholesale market) | 56,052 | Business |
| 超市 (Supermarket) | 54,615 | Business |
| 小学 (Elementary school) | 51,989 | Place |
| 大厅 (Hall) | 51,860 | Place |
| 购物中心 (Shopping mall) | 50,547 | Business |
| 加油站 (Gas station) | 49,748 | Business |
| 礼品店 (Gift shop) | 49,171 | Business |
| 公厕 (Public toilet) | 48,633 | Place |
| 健康食品专卖店 (Health food store) | 47,833 | Business |
| 停车场 (Parking lot) | 46,831 | Place |
| 床垫专卖店 (Mattress store) | 46,304 | Business |
| 酒吧 (Bar) | 46,145 | Business |
| 银行 (Bank) | 46,106 | Business |
| 物流服务 (Logistics service) | 45,969 | Business |
| 综合医院 (General hospital) | 44,160 | Business |
| 门窗店 (Door / window store) | 42,990 | Business |
| 运输服务 (Transportation service) | 39,918 | Business |
| 纹身店 (Tattoo shop) | 39,617 | Business |
A smaller localization gap than Japan's
China's category data has the same kind of taxonomy crack Japan's did, just narrower: 2 of the top 50 categories are raw English tokens — atm (60,152 listings) and spa (59,349) — instead of localized Chinese labels. Every other category in the top 50 is a proper Chinese term. Japan's equivalent cut had 4 untranslated tokens (shinto_shrine, buddhist_temple, place_of_worship, atm) covering a larger combined share. This is a Google-side taxonomy artifact, not a Crawlora extraction bug in either market — worth checking for explicitly if you're building a cross-language category filter on this dataset, since assuming every market's category field is uniformly localized will silently miss whatever fraction leaks through in raw English.
What this means for anyone scraping or filtering Google Maps data
The practical rule from comparing these two markets: don't assume a "top categories" ranking means "top businesses" without checking market by market. Japan needed an aggressive point-of-interest exclusion list (shrines, temples, parks, apartment complexes, community centers) before its category ranking meant anything about businesses specifically. China's raw, unfiltered top 50 is already 92.4% commercial — the same denylist barely moves the needle here. A single global POI-filtering rule tuned on one market's data will over-correct on the other.
If you need this filtering already handled per market, Crawlora's Google Maps API exposes the same category facet used in this post, filterable by exact value, so you can build a market-specific allowlist or denylist directly against live data rather than assuming one filter generalizes everywhere.
Methodology & data
Every figure in this post comes from a live facet aggregation (facet=category) against Crawlora's google-map-businesses dataset, filtered to country=China, pulled July 20, 2026, covering the 50 largest categories by listing count. "Business" vs. "Place" classification is a manual read of each category's real-world nature (a judgment call, not a field in the dataset), following the same classification convention as the companion Japan post; a small number of categories genuinely ambiguous between the two (kindergartens, hospitals, corporate offices) were classified as businesses since they're commercial service providers, not civic/public infrastructure. The uncategorized bucket (289,459 listings with no category value) is shown separately and excluded from both the business and place totals.
This is a point-in-time crawl snapshot, not an authoritative registry of business types in China. We publish aggregate rollups only; no individual business record from this dataset is exposed in this post.
Query China's Google Maps category data directly
7.3M+ China listings across every category — filterable by exact category value, province, and rating, over one REST API.
Frequently asked questions
What kind of businesses does Google Maps index in China?
Overwhelmingly commercial ones. 9 of China's 10 largest Google Maps categories by listing count are genuine businesses — Chinese restaurants (549,713 listings), fast food, shops, convenience stores, beauty salons, clothing stores, hotels, and pharmacies. The only non-business entry in the top 10 is an uncategorized bucket.
How does China's Google Maps category mix compare to Japan's?
It's close to the opposite. Only 7.6% of listings across China's top 50 categories are non-commercial places (government agencies, ATMs, schools, parking lots) — versus 40.1% in Japan, where shrines, temples, historic sites, and parks dominate the top of the ranking.
Are there localization gaps in Google Maps' China category data?
Yes, a smaller version of the same gap found in Japan's data: 2 of China's top 50 categories (atm, spa) are untranslated English tokens rather than localized Chinese labels. Japan's equivalent cut had 4 untranslated tokens.
What's the largest single Google Maps category in China?
Chinese restaurants (中餐馆), with 549,713 listings — more than 3x the next-largest identifiable category (fast food restaurants, 164,689).