Usage Guide
A practical walkthrough of the Global Location Data Platform (GLP): what it shows, how to read it, and what its limits honestly are.
What this platform is (and is NOT)
Everything you see is verifiable: each place carries its source, version, license, and raw record so you can trace it back to the original dataset.
Getting started — 3 steps
- Choose a location. Type a city or address into the search box (it geocodes and flies the map there), tap one of the quick-city buttons (Riyadh, Jeddah, Makkah, Cairo, Dubai, London, New York), or simply click anywhere on Earth.
- Pick a radius. Use the pills — 500 m, 1 km, 2 km, 5 km, or 10 km — to set how far around the point to look. The dashed reference rings show an ~800 m walk and a ~4 km drive for scale.
- Read the results. Nearby businesses and places load automatically, grouped into 16 category cards with counts, plus a paginated place list. Click any place for its raw details.
Reading the map & layers
Layer chips
Toggle what appears on the map without changing your search:
- Thematic layers: businesses, community places, transit, parking, hotels, healthcare, food.
- Context layers: nearby cities (with population), airports (with IATA codes), and the country outline.
Basemaps
Switch between Streets, Satellite, and Simple styles. Each has tile fallbacks, so the map keeps rendering if one tile server is slow or unreachable.
Search box
Free-text queries are geocoded through the platform’s explore API; the map animates to the match and reloads places within your chosen radius.
Understanding categories & the “Other” bucket
Places are grouped into 16 readable categories: Food & Drink, Grocery, Retail, Healthcare, Beauty & Wellness, Hotels, Education, Fitness, Automotive, Finance & Legal, Entertainment & Culture, Professional Services, Public & Civic, Places of Worship, Parks & Outdoors, and Other.
“Other” is not a dumping ground you have to trust blindly. It is clickable: open it and you see the raw source labels (e.g. original OSM/Scheme tags) behind each grouped place, so you can judge the grouping yourself. Category grouping is a presentation aid — the underlying label is always preserved.
The three honest empty states
When a search returns nothing, GLP tells you why instead of showing a blank map. Three distinct messages exist, and the difference matters:
| Message | What it means |
|---|---|
| “Business information isn’t currently available for this area” | The point falls outside the footprint of any ingested source — no dataset covers that exact spot. The emptiness is about source coverage, not about the place being uninteresting. |
| “No places were returned within this area” | Sources do cover the region, and the query ran fine — there genuinely are no recorded places inside your radius. Try a larger radius. |
| “Sources for this region are being added” | This region is not yet ingested into the platform. Data may exist upstream; it simply isn’t loaded here yet. |
GLP never guesses or fills gaps with synthetic data. An empty result is information.
Detail panel & provenance
Click a place in the list or on the map to open its detail overlay. It shows raw fields only — name, category, address, phone, website, opening hours, brand, and whatever else the source recorded. Nothing is embellished or inferred.
Expand “Advanced data” for full provenance:
- Source — which dataset the record came from (e.g. Overture, OSM, GeoNames).
- Version — the snapshot/release of that dataset, so you know how current it is.
- License — the terms attached to the data; check these before redistributing or exporting.
- Record ID & raw JSON — the exact original record, for verification.
Site Analysis — what surrounds a point
On the home page, after you press Analyze this location, the chips under the area title give a one-glance read of an area instead of scrolling raw lists. Pick a location and a radius, then the panel answers: how many people, what kind of land, how dense with businesses. Use Compare sites to see those same measurements side by side for 2 or 3 locations — factual differences only, no ranking.
- Population in range — an estimate of residents within your radius, summed from WorldPop 100 m cells. It is a reference figure, not a census.
- Land cover — the dominant surface type (built-up, cropland, bare, water…) from ESA WorldCover, with the share of each class in the area.
- Business density — how many places per km², counted directly from the places already on the map. A raw count, never a demand forecast.
- Buildings / Night activity / Climate — shown only where a stored source actually covers your point.
Every number has an ⓘ button that opens its source, version, license, and how it was computed. Income, employment and rent are not available yet — no source in the catalog carries them, so the panel says so plainly instead of inventing a figure.
Compare sites: after selecting Site A (a drawn area or a point), press “Compare sites”, then use ✏️ Draw area to draw Site B — Site A stays on the map (blue). Draw Site C too if you want (orange B, green C stay visible together), or skip it. Then press Analyze Sites. Each site is analyzed independently from its own drawn geometry and shown in its own A / B / C section with a compact raw-numbers summary table — no winner, no score, no ranking. The judgement stays yours.
Co-location & White Space are yours to define: before analyzing, you can add the exact categories you want checked around the location (e.g. Gym, Supermarket, Hotel). The platform never assumes which categories are complementary or where a “white space” is — it only reports what was observed for the categories you entered, with the nearest match, the source, and the query used. “Not observed” means no record was found in the loaded sources — it is not proof of absence.
Exporting data
You can download the places currently visible in your search as GeoJSON, CSV, or GeoParquet.
- Exports reflect what is on screen for your current center + radius — adjust the search first if you want a different set.
- Exports are license-aware: each source carries its own open-data license (e.g. ODbL for OSM-derived data). You are responsible for honoring attribution and share-alike terms when you reuse the data.
Coverage
The platform ingests sources covering roughly 237 of 242 countries and territories. Only Antarctica and a handful of tiny remote islands have no data. Coverage of places within a covered country still varies — dense cities have far more records than rural areas, and completeness depends on the underlying open sources, not on GLP.
FAQ
Does GLP rate or rank areas for me?
No. GLP shows raw facts only. It has no scores, no “best neighborhood” logic, and no predictions. Judgements are yours to make.
Why does a shop I know exist not appear on the map?
Either no ingested source has recorded it, it sits outside your radius, or its category isn’t in the loaded layers. Empty areas come with an explicit message explaining which case applies.
How fresh is the data?
Each record’s “Advanced data” panel shows the source and version (snapshot) it came from. Open datasets update on their own cadence — the version stamp tells you exactly which release you are looking at.
Can I use exported data commercially?
It depends on each source’s license, which is shown per record and attached to exports. Many open licenses (like ODbL) allow commercial use but require attribution and share-alike conditions. Check before you publish.
Can I analyze a list of locations at once?
Yes — the Bulk analysis page takes a CSV list of sites and returns one raw per-site observation table row each. It is independent from the map analysis and runs one Google collection per site (up to 10 sites).
Why are there dashed circles on the map?
They are scale references only: ~800 m (a typical walk) and ~4 km (a short drive). They don’t affect your search — the solid radius circle does.
The map tiles look slow or blank — is the data gone?
No. Basemaps have tile fallbacks; if a tile server is unreachable you may see a plain background while place data still loads. Try another basemap style.
Does GLP store my searches?
GLP is a viewer over public open data; the guide page itself stores nothing. Treat the platform as a read window, not a personal database.