Fresh, accurate, & precise data about places

What is SafeGraph?

SafeGraph tracks physical places around the world, capturing precise geocodes and context like place type, opening dates, and brand affiliation for millions of points of interest. On Dewey, SafeGraph also publishes anonymized transaction data tied to those same places, so researchers can see not just where a location is, but how much people spend there. Both datasets refresh monthly and share Placekey as an identifier, so they join together cleanly.

‍

What academic researchers should know about SafeGraph places and spend data

Global Places covers points of interest worldwide with precise geocodes, place type, and open and close history, refreshed monthly since 2019. Spend Patterns covers United States points of interest only, aggregating anonymized credit and debit transactions each month back to January 2019, including total spend, transaction counts, and demographic breakdowns by income and home city. Every Spend record ties back to a specific point of interest through Placekey, so transaction activity inherits the full context of the place it happened at. Missing rows in Spend represent suppressed data for privacy, not zero activity, which matters for sample construction.

‍

Why academic researchers choose SafeGraph on Dewey

SafeGraph earned its academic reputation by opening data to researchers in 2020, and six years of published work speaks for itself: more than 700 papers across urban planning, economics, public health, and finance. Every point of interest carries a persistent Placekey, precise coordinates, industry code, and brand affiliation, so a specific coffee shop or urgent care clinic can be tracked over time and compared against thousands of similar establishments nationwide. That structure pairs naturally with other data on Dewey. Advan Research picked up SafeGraph's foot traffic movement methodology after SafeGraph stopped publishing Patterns data in 2022, so researchers who need movement data alongside places and spend can combine both. ATTOM adds property level detail for work that touches real estate, ClimateCheck adds physical climate risk ratings for the same points of interest, and Similarweb extends the same establishments into digital behavior for omnichannel research.

‍

SafeGraph academic research ideas and use cases

Urban planning and transportation research is where SafeGraph shows up most often on Dewey, covering downtown recovery, neighborhood change, and how services are distributed across cities. Researchers at Texas A&M, the University of South Florida, Grand Valley State University, and Emory University built a graph neural network on SafeGraph Places data to model how different types of urban locations recover after a disruption, using the point of interest structure to link establishments by type, proximity, and shared visitors.

‍

Public health and mobility questions are a close second. A University of Florida study paired SafeGraph Places data with search and social media trends to measure how the pandemic changed travel distances and search interest around major landmarks, using the places data to anchor exactly where visitors were headed.

‍

Finance and real estate researchers use the same places data to study how development changes a neighborhood. A University of Pennsylvania paper drew on SafeGraph alongside other micro level data to test whether new NFL stadiums produce lasting agglomeration effects nearby, tracking housing prices and business activity before and after construction.

‍

Spend data opens up a different set of questions, ones that need transaction level detail rather than just location. MIT researchers used SafeGraph Spend data across more than 4,000 EV charging stations and 140,000 California businesses to show that a single new charging station raised nearby spending by 1.4 percent, worth an estimated $6.7 million in local economic activity concentrated in food, entertainment, and retail.

‍

Beyond these published examples, Spend supports research designs that need a clean treated and control group of businesses: what happens to nearby retail spending when a competitor opens, a transit stop closes, or a tax holiday takes effect. Because every record ties to an individually identified point of interest rather than a zip code average, researchers can build the event study and difference in differences designs that applied microeconomics depends on.

‍