Matching SafeGraph businesses to other business databases using name / address

Hi, we are matching SafeGraph businesses to other business databases using name / address in order to benchmark and obtain additional information about these businesses. The quality of the match depends crucially on standardizing name and address prior to matching. 3 questions:

(1) @ross_epstein_safegraph What is the source of Safegraph’s business name / address / lat-lon information? Has it been standardized in any way and if so how?

(2) We have been using Google places API to standardize name / addresses in our outside databases and to obtain lat-lon at the same time. But this is costly ($17 per 1000 API calls) given that we want to standarize millions of name / addresses. Can anyone with large Google API capacity help (or know someone at Google to ask for a “free pass” :slightly_smiling_face: )? Can SmartyStreet help (@Jeffrey_Duncan_SmartyStreets_LLC kindly offered such a service in March on this channel)? Many thanks!!

(3) Alternatives to standardizing name and addresses and then using a match algorithm (I know of a project that uses the web search API to obtain URLs of businesses and match in that way; but that also requires a paid subscription and not clear it’s better given that SafeGraph comes with lat-lon info)

Sorry for the long post and many thanks for your help!!

@Andre_Kurmann_Drexel_University we have a free service called Match that will do this matching for you:

highly recommend

Smarty streets can definitely help you standardize addresses at scale (which will increase your match rate in the SafeGraph matching service).

Thanks @Ryan_Fox_Squire_SafeGraph. We’ve been using your match service and it’s quite useful. Unsurprisingly, some of the match results differ depending on whether we use raw business names / address (as users enter them into the database we want to match) or Google names / addresses (obtained by running raw names / addresses through Google API first). In addition, running Google API provides lat-lon which helps improving the match results.

That’s why we’d like to run all of our database through Google Places API first (or another service) before matching with SG.

Getting help from SmartyStreets on this would be fantastic. @Jeffrey_Duncan_SmartyStreets_LLC, please let us know if this is possible – many thanks!