Do you believe the piece linked in the first paragraph[0] describes an outlier in terms of data collection? It shows how location data from a single low-lying location tracking company had enough pings to reliably pinpoint home and work locations of individuals, enough to de-anonymize and interview them. This does not seem like sparse data.

[0] https://www.nytimes.com/interactive/2019/12/19/opinion/locat... (discussion https://news.ycombinator.com/item?id=21833718)