Proximity Dataset

A segmentation of neighborhoods based on the type and velocity of social activity in the area.

data Overview

Reveal an area's activity through people’s geotagged posts.

The Proximity dataset scores block groups across 72 behavioral segments. Sources include public geotagged posts from Twitter, Instagram, Facebook, Flickr, Foursquare, Meetup, and more.


Social users


Data points


Block groups


Unique segments

What is Geosocial proximity?

Location-based social media data

The Geosocial Proximity dataset is a classification of areas based on what people are doing and saying there. It originates from where the post was tagged. By combining traditional demographics in trade areas with social behaviors in close proximity, retailers and restaurants can reveal hidden location characteristics that drive business success.

Partner Integration

Activate the Geosocial Proximity dataset in a mapping and analytics platform from one of our many partners.

Feature Service

Esri users can leverage a pre-built feature service and suite of plug-n-play reports and infographics.

Raw Data

Use the raw data in CSV or GeoJSON format within your statistical analysis software such as Python, R, Tableau, PowerBI, and Alteryx.

Key Features

Capture the heartbeat of an area

Demographics fall short especially when the visitors to a trade area don’t live in that location. Geosocial picks up the volume and activity of a location to capture the full picture of area characteristics.

Sharper predictive models

Enhance predictive models by differentiating demographically similar locations by social behavior. The different behaviors in an area can often explain high and low performing stores that are otherwise demographically the same.


Unlike Census data which can lag up to 10 years, Geosocial Proximity is updated quarterly showing the current and historical trajectory of an area. Keep models up to date with current behavior and know the trajectory of an area to make smart real estate acquisitions.

How it works

An overview of how this dataset was built.



People geotag posts about what they are doing or experiencing in a location.



We clean posts and aggregate at the block group to analyze text and hashtags.



Using machine learning we cluster mathematically related words into segments.



We score block groups as percentiles 0 - 100 vs. the nation. Above 50 is above average.

Ready to get started?

Reach out to our sales team to see a personalized demo for your use case. Or explore the Proximity taxonomy to see the data for yourself.