Overview¶

Spatial.ai's GeoWeb dataset summarizes, at a geographic level, web-browsing behaviors across more than one thousand unique segments. In general, this dataset answers the question: what types of websites and online behaviors do the people who live here participate in more than the average American?

However, these segments get extremely specific. Segments like Thrift Store Shoppers, Doctors, Hands-On Parents, Sleep Disorder Researchers, and Vegans are just a few examples from the 1000+ variables included in this dataset.

Deprecation¶

As of October 2022, the GeoWeb dataset will no longer receive updates. Shifts in the underlying collection methods and strategy of Dstillery, the provider which allowed us to generate this dataset, have necessitated the dataset's discontinuation.

Partially as a response to this discontinuation, Spatial.ai developed the FollowGraph dataset. FollowGraph is based on social media data rather than web-browsing behavior, but is structured identically to GeoWeb, contains similar variables, and answers similar questions about peoples' interests in activities, celebrities, media, brands, and more. We recommend exploring FollowGraph as an alternative to GeoWeb.

The standard GeoWeb dataset is based on 2 years of web-browsing data. While updates are no longer forthcoming, the data will remain extremely useful and insightful for years to come. Do not hesitate to reach out to our team for a sample if you would like to explore licensing the dataset. Multiple studies have shown location datasets provide significant predictive power for years, even decades, after their original creation.

Why use GeoWeb?¶

The strength of the GeoWeb dataset is in its wide breadth and unique source. Web browsing behavior is a data source unlike traditional datasets that allows you uncover relationships across a host of different interests that may be correlated with geographic outcomes. Similar to the Proximity Dataset, this data source is generated independently of demographic data. Therefore, it captures different signal that can vastly improve predictive models.

Sample Data¶

You can find sample data for this data at this link.

You can find a shareable one-pager summarizing GeoWeb at this link.

Coverage¶

This dataset is available across the entire USA at the census block group level.

Changelog¶

You can view our public changelog for the GeoWeb dataset at this at this link.