Serious question: Is it too early for pumpkins? 🤔 Welcome to the September 2025 release notes! 🍂

Places Highlights

137k duplicate places removed from US Data 👯‍♀️
79 new brands across 40 countries 🎉
Precise Geometry for US/CA Shopping Malls 🛍

Places Growth

This month, SG Places has a grand total of 75,907,475, including POI with or without geometry, closed POI, and parking lots. This is a net decrease of 67,917 places from last month. Outside of the US, most countries realized modest coverage gains through regular source updates and new brand additions.

Of course, you can always visit our Places Summary Stats to find more details on our continued growth.

US Duplicate Places

In the US, we prioritized precision over coverage efforts this month and especially focused on removing obvious duplicate place examples. These primarily include "branded" ↔️ "non-branded" places at the same street_address pairs, same physician at same street_address pairs, and otherwise same "non-branded" place at same street_address pairs. As a result, we identified and removed ~137k duplicate POIs from the September release.

We pride ourselves on the ability to merge like places, and while our deduplication model does well balancing over-merging (removing real, distinct places) with under-merging (retaining true duplicates), occasional audits are necessary, and our model learns from the heuristics used in such audits. Notice any more obvious dupes? Let us know! Eradicating duplicates from our data is a never ending journey, and we appreciate your feedback 😄.

Brands

This month, we added a grand total of 79 brands across 40 countries including:

Sberbank (SG_BRAND_f4603d3f14682b6c) with 11,248 POIs 🏦
Axis Bank (SG_BRAND_76cb4cd14fc99701) with 5,782 POIs 🏦
Strickland Brothers 10 Minute Oil Change (SG_BRAND_bffffeb6ec081912) with 215 POIs 🚗 🛢️
Boucheron (SG_BRAND_d8b164686688abff) with 84 POIs 💍
Simply Southern(SG_BRAND_eb21ffb2714110fd) 32 POIs 👕 👗

👀 Are we missing a brand or country? 👀 Please let us know here

Brand Openings and Closings

We rely on POI metadata to track store openings and closings, and we are especially interested in understanding open/close dates for branded POIs. It can take more than a month to infer open/close dates, so we report brand open/close metrics on a one month delay.
In this release, we flagged 1,894 brands with at least one store closure in July 2025 and 2,278 brands with at least one store opening in July 2025. Learn more about our open/close columns here.

Drops ⬇️

We are ingesting many sources and due to source changes and processing changes, Placekeys do drop over time. In this release, we dropped 940,755 Placekeys (65,117 branded and 875,638 non-branded).

Precise Geometry for US/CA Shopping Malls

We know... we sound like a broken record announcing "more improvements" for mall geometry and spatial hierarchy. Over the past several months, we've adjusted our parent ↔️ child spatial hierarchy rules to infer more direct mall ↔️ tenant relationships, sourced new geometry for problematic mall examples, removed duplicate malls, etc. All of these efforts lead to marginal improvements, but mall issues continued to persist at scale. To put a painkiller to this headache once and for all, we are thrilled to introduce brand new geometry for US/CA shopping malls! ✨

By the numbers

+489 mall POIs across US and CA (naics_code = 531120) including recently opened malls 📈
2,155 US/CA malls flipped polygon_class from "SHARED_POLYGON" to "OWNED_POLYGON" 📈
A whopping 95% of the 46,469 previously existing US/CA malls have a new polygon_wkt 💥 ‼️

Many of the polygon_wkt changes are minor (< 5% change in wkt_area_sq_meters), but others are big, intentional changes that address the primary issue: displaced mall geometry. This issue results in polygons for the incorrect/neighboring mall which can lead to overlapping mall polygons (creating a spatial hierarchy mess) or missing mall polygons altogether.

September-2025 Release Notes