No, both home and work visits are filtered out of our first party visits. To calculate a user’s work, Foursquare looks at at least 56 days worth of visit data, focusing on the visits of long duration (at least two hours). Foursquare uses these visits to build a set of places where each user has spent a large majority of time. We then assign these location clusters as home/work, and tag visits to these locations accordingly. Foursquare differentiates home from work based on the proportion of the time of day the user has spent at said location. Typically if the user visits the location during business hours, then we tag that location as work. If the user spends long periods of time in a place outside of business hours, then we tag that location as home.
In rare cases, Foursquare data may show a visit to a store location even when the store is not open. For example, me may capture some visits to Chick-fil-A on Sundays due to individuals who are standing in close proximity of the store for several minutes, resulting in a visit being recorded (inaccurate snaps). We’re consistently improving our clustering methodology to increase our visit precision and reduce false positives.
The visit may be missing a normalization weight (SAG score) because we have not been able to infer the user’s demographics (state, age, and gender), which are required inputs for the SAG score. Once we can infer the user’s demographics, we attach a normalization weight to their go-forward and historical visits. We are exploring ways to potentially remove visits without normalization weights. In the meantime, please disregard these visits in the dataset.
Updated 7 months ago