Home

The Problem

The Solution

Nuts and Bolts

Applications

Advantages

Patent

Contact

Blurring – residential and nonresidential areas

For residential areas, we can control the extent of blurring because we have good data on population density; blurring to the census tract level might do the trick, and the census tells us what geographic distance we have to go to in order to blur that much.

Blurring will be less precise in nonresidential areas if we don’t know the densities there. For example, when someone is in a non-retail commercial sector, we may not know whether that one building has enough people to provide anonymity, or whether we need to blur to the block, or larger unit.

What is needed is a map containing data on the density of people in each area at representative times throughout the day. Happily, it is relatively easy to collect this density data without significant privacy risk. How? The privacy risk comes primarily from tracks which show someone going from X, to Y, and, then, to Z. But, to get the density data, we don't need to link someone's being in X, with that person's subsequent travel to Y and Z. So, we can collect density data without too much risk by using a “snapshot” method, rather than a tracking method; we'd simply record device locations every sample interval without including in the dataset a unique identifier which would otherwise make it possible to tell which phone was which (and where each phone was previously).

Once the density data is obtained, it can be used to blur (in real time) locations in nonpublic, nonresidential areas.

Back to "Misc. Details on the Method"

What can you do with this data?