Note: As of 2021, there is no longer a free tier for using the Google Nearby Places API, so this method cannot be used for free anymore.

argo

this project can help you get lots more places from the Google Maps places API

getting started

install project dependencies with:

npm install

You'll also need a Google Maps API key. Instructions to get one can be found here.

running the script

this script is written to run for restaurants, but could run for any place types that Google lets you search for.

Create a config.json in the root of this folder, copied from the config.example.json (also in the project root).

Provide the Google Maps API key, the place type, the radius you want to search in (the default from the example is a good radius), and the latitude and longitude of the grid you want to get restaurants for.

The lat / long of the grid will be in the format of [35.0, 35.18], where the lower number comes first. Negatives apply for southern hemisphere latitude and for western longitude (the equator and the prime meridian being zero, respectively).

The example coordinates are for a square that roughly covers Chattanooga, TN.

why? this seems strange

Google limits you to 60 responses per api call, so you can't just get back all of the places of a given place type in a large search area (of a circle with a radius of 10km, say).

You can, however, search with a small enough search area that you can get back all of the results for that smaller circle.

This repo operates by using a small enough search area (a small circle with a radius of 2km), and doing many searches over the larger search area. Instead of doing one search for a in a 10km radius circle, we do 100 smaller searches in a 10km x 10km square with overlapping search area. That way, we get all of the places in the square. See fig. 1 below:

This sweeping of the grid is similar to how they found the Titanic; Like Robert Ballard, we sweep back and forth across the search area to find all the things we're looking for.

Want to read about this project in article form? I wrote about a more in depth explainer of it on my blog: https://samthomas.io/argo/