Structured attributes for places in a supported country.
Define the geography relevant to your question.
Explore counts, categories, and other supported attributes.
At A Glance
02 / What's Included
Included
Not Included
primary_type is a single category; types can contain several. Type counts can overlap. Use business_status to select operational businesses and user_rating_count to contextualize ratings. brand_ids links to supported brand data; snapshot_date is available in historical views.
Sample Place Types
Sample Attributes
03 / Data Source
Key Facts
Combine permitted aggregate results with your own data in BigQuery. Query controls and applicable license terms govern access and export; subscription does not provide unrestricted raw-table extraction.
Curated place attributes for the subscribed country.
Monthly refresh and documented historical snapshots.
Combine permitted results with your own BigQuery data.
Use supported SQL or count functions. Downstream analysis can use Looker Studio or a separately validated model.
04 / Extend With Maps Platform
Pairs Well With
05 / Example Queries
Operational restaurants by H3 cell within a 2-mile circle
SELECT * FROM `YOUR_PROJECT.places_insights___us.PLACES_COUNT_PER_H3`( JSON_OBJECT( 'geography', ST_BUFFER(ST_GEOGPOINT(-97.7431, 30.2672), 3218), 'h3_resolution', 8, 'types', ['restaurant'], 'business_status', ['OPERATIONAL'] ) );
Example SQL · No request executed
Illustrative aggregates within a 2-mile circle
SELECT WITH AGGREGATION_THRESHOLD
primary_type, COUNT(id) AS place_count
FROM `YOUR_PROJECT.places_insights___us.places`
WHERE ST_DWITHIN(point, ST_GEOGPOINT(-97.7431, 30.2672), 3218)
AND business_status = 'OPERATIONAL'
AND primary_type IN ('cafe', 'restaurant', 'fast_food_restaurant')
GROUP BY primary_type
ORDER BY place_count DESC
Illustrative output · No request executed
| primary_type | place_count |
|---|---|
| restaurant | 42 |
| cafe | 18 |
| fast_food_restaurant | 11 |
Price categories are grouped; AVG ignores missing ratings
SELECT WITH AGGREGATION_THRESHOLD price_level, COUNT(id) AS place_count, ROUND(AVG(rating), 2) AS avg_rating FROM `YOUR_PROJECT.places_insights___us.places` WHERE ST_DWITHIN(point, ST_GEOGPOINT(-97.7431, 30.2672), 3218) AND business_status = 'OPERATIONAL' AND primary_type = 'restaurant' AND price_level IS NOT NULL GROUP BY price_level ORDER BY place_count DESC
Illustrative output · No request executed
| price_level | place_count | avg_rating |
|---|---|---|
| PRICE_LEVEL_MODERATE | 24 | 4.32 |
| PRICE_LEVEL_INEXPENSIVE | 10 | 4.10 |
| PRICE_LEVEL_EXPENSIVE | 8 | 4.41 |
06 / Use Cases And Industries
The published Tokyo and Yokkaichi study combined parcel prices with amenity counts within 500 meters.
The study reported different associations by city: subway access in Tokyo and universities in Yokkaichi featured among the positive relationships.
These findings describe correlation. They do not establish causation or transferable price premiums.
07 / Privacy And Delivery
Direct Query
Count Functions
Aggregations over protected tables; threshold ≥ 5; individual Place IDs are not returned.
Exact counts, including zero; eligible functions can return limited Place IDs. Minimum area and other constraints apply.
Use PLACES_COUNT_PER_H3 for density maps. The get-started guide also lists PLACES_COUNT_V2 and PLACES_COUNT_PER_TYPE_V2 for batch geographies, and PLACES_COUNT_CHANGE for monthly comparisons. Check their launch stages and availability before adopting them.
08 / Evaluation
| Evaluation Area | Places Insights | Ask Of Another Offering |
|---|---|---|
| Source | Google place information | What are the source and coverage limits? |
| Access | BigQuery SQL and supported count functions | Which interfaces and export controls apply? |
| Refresh | Monthly dataset refresh | What refresh and historical coverage are documented? |
| Cost | Dataset access plus applicable BigQuery usage | Are there subscription, query, or export charges? |
| Schema | Documented fields and country availability | How consistent are fields across geographies? |
| Brands | Separate supported brand data | Which brand definitions and countries are covered? |
This is an evaluation checklist, not a competitive ranking. Record dated sources before making provider-specific comparisons.
Map counts with PLACES_COUNT_PER_H3, select areas of interest, and inspect returned sample_place_ids with Place Details. Compare the mapped samples with your filters and record discrepancies. A sample check does not prove complete coverage.
Move from regional screening to local H3 cells. Combine complementary place types and opening hours with documented weights, then overlay competitors. A weighted suitability score expresses your criteria; validate the shortlist against access, lease costs, and local conditions.
Join your measured outcomes to nearby POI counts, train a model, and evaluate held-out predictions before applying ML.PREDICT. Report MAE and R²: neither establishes causation, and MAE is an average error, not a prediction interval.
The linked notebook uses synthetic store scores. Its fit does not establish real-world accuracy. Our implementation recommendation: match training and scoring catchments, filters, and dates; 500-meter circles and H3 cell counts are not interchangeable. Test geographic or temporal holdouts and justify any outcome-based exclusions.
Normalize CRM accounts to Place IDs, exclude matches, and check unresolved or duplicate records. Count functions return at most 250 sample IDs per response element; a capped sample is not a complete territory list. Refine filters or geography within supported limits and deduplicate results.
Use Place IDs for recurring matching and fetch current details when needed. Apply the Places API storage, attribution, and billing requirements to those lookups.
09 / Coverage And Pricing
Coverage
Cost
50 separate country datasets make up the full paid catalog — each subscribed individually as its own BigQuery listing