---
title: "Home & work inference — USA"
description: "Estimating the likely home and work areas of visitor groups, always at area level and never an address, to enable resident / worker / visitor breakdowns, trade areas, and visitor-only scoping for Visit Length."
slug: "/docs/datasets/usa/methodology/home-and-work"
last_updated: "2026-08-31"
related_to: ["extrapolation", "privacy-approach"]
tags: ["methodology", "home-and-work", "usa"]
country: "USA"
applies_to: ["gps", "telco"]
---


# Home & work inference

A likely home area and work area are inferred from activity patterns (home areas from sustained night-time presence, work areas from sustained day-time presence). "Home" here always means a statistical area: a neighborhood-sized zone, never an address, a building, or a household. These area anchors feed the resident / worker / visitor split in foot traffic, the visitor-only scope of visit length, and, most importantly, the home and work dimensions of trade areas.

## What "home" means, and what it can never reveal

Pine59 datasets do not contain, and cannot be used to derive, where any individual lives or works. "Home area" is an aggregated statistical concept, protected by several complementary layers:

- **Area level only.** Home and work are inferred as coarse areas (a census block group, a postal area, a statistical tract), never a street address, a building, or a coordinate.
- **Aggregation only.** The inference is an intermediate input to aggregate statistics. What is published is always a group-level figure, such as "12% of this store's visitors have their home area in this ZIP code". No individual-level record exists in any output.
- **Aggregation at the source.** Depending on the signal, aggregation begins before the data reaches Pine59, so individual-level traces are not part of our pipeline.
- **Minimum-cohort suppression.** Any home-area statistic representing fewer than a minimum number of visitors is suppressed entirely (see *K filtering*), so small groups can never be singled out.
- **Rounding.** Published figures are rounded, so no output implies a precise individual observation.

Together these mean the question "where does this person live?" is not answerable from Pine59 data, by us or by anyone else. The only question the data answers is "roughly where does this *crowd* come from?", which is exactly what trade areas and visitor profiles are for.

Applies in USA because it is served by a gps or telco signal. See [Signal](/docs/datasets/usa/supply.md).

## Related

- [Extrapolation](/docs/datasets/usa/methodology/extrapolation.md) — Scaling the observed sample of devices up to population-level visit estimates, so counts reflect real-world volume rather than sample size.
- [Privacy approach](/docs/datasets/usa/methodology/privacy-approach.md) — Pine59 products are built for statistical use only, and no output can be linked to an individual person.