All Our Data

Final Results

All Our Data

Advisor: Jeff Ondich, W/S

Video introduction

Background

Pretty much anyone with the funds, equipment, time, and inclination to participate in the modern networked world has had an experience like this. The other day, I was thinking it was probably time to buy a new pair of hiking boots. Later that day, boot ads started showing up in my Facebook and Twitter feeds. How did they read my mind? I dunno, maybe I said “I need new boots” aloud and Siri grabbed the sound. Or maybe I tweeted something about hiking in a state park, or maybe I mentioned dog walks in an email. Or maybe I bought these boots exactly two years ago, and REI told the advertising cartels it was time for me to buy another pair. However it happened, it was creepy. But yeah, especially sitting at home (except for hikes in the woods) on state-wide lockdown, I considered clicking on the ad—those boots looked awesome.

There is a huge data brokerage industry dedicated to the construction of mathematical models of each of us, the better to support targeted advertisements, targeted political persuasion, and other forms of manipulation and discrimination for profit. This industry, described in alarming detail by Shoshanna Zuboff in The Age of Surveillance Capitalism, is often talked about, but its mechanisms are mostly hidden behind opaque corporate names and mushy language. The extent to which the public knows anything at all about this industry in the US is largely due to a 2018 Vermont law requiring registration

In recent years, journalists and researchers have tried to learn more about the type of data collected about us, and how that data is extracted, stored, sold, and exploited. For example, journalist Kashmir Hill's 2019 article I Got Access to My Secret Consumer Score. Now You Can Get Yours, Too demonstrates, among other things, the layers of bureaucracy hiding the data industry, as well as its lack of governmental regulation. (Interested in this stuff? You should read everything Hill has written in the last couple years. She's enormously creative in how she extracts knowledge out of an industry that is opaque by design.)

The goal of this comps project will be to learn as much as we can about the data industry and the data it collects. What data do they collect about us and how is it structured? Who has it? How did they get it? How is it used and abused? Does their use of our data yield benefits for us? Does it hurt us (or, maybe more to the point, who does it hurt?). And so on.

This is an open-ended and exploratory comps project, which will require creative thinking, an experimental mindset, and a tolerance of uncertainty. But it also has the potential to give us deeper knowledge than is currently available of how the data industry studies, models, and manipulates us.

The project

Here are the kinds of activities you will pursue during this project.

Read up on the data economy and projects like Hill's and others to figure out what other people have done.
Do a literature search on data ethics and data privacy to learn the arguments for various approaches to privacy regulation.
Collect as much of each team member's personal data as possible from as many sources as we can. (We will keep our data private from each other, but as we learn how to get data from various sources, we will all be able to study our own data to draw general conclusions and perform experiments.)
Study and describe the structure and content of the various data sources.
Theorize how the data was acquired by the party(ies) in question, and try to construct and run tests of your theories.
See if you can find overlap in the datasets suggesting that one party acquired the data from another.
For each individual data source, assuming a corporation or government possessed that data for millions of people, assemble a list of potential applications to which they could put the data.
Same thing, but now imagine combining the various data sources, plus any publicly available datasets that might be available. What applications are possible now?
Keeping our studies of data ethics in mind, devise and perform experiments to see whether you can manipulate the uses of your own data indirectly (e.g. can you force Facebook to show you ads for teddy bears or water skis?)
etc. (You'll have some great ideas I haven't thought of.)

Deliverables will include:

A bibliography.
A paper summarizing your investigations and conclusions (presented online, since it's likely to be packed with links). This paper will be intended for a general audience (think the New York Times or The Atlantic).
Any code you used to automate your experiments, do statistical analysis, etc.
Other deliverables suitable to whatever activities this adventure leads us into.

Prerequisites

Curiosity, plus a tolerance for ambiguity and uncertainty.

References

The Secretive Company That Might End Privacy as We Know It, Kashmir Hill, New York Times, 18 Jan 2020

I Got Access to My Secret Consumer Score. Now You Can Get Yours, Too, Kashmir Hill, New York Times, 4 Nov 2019

H.764 (Act 171) of the Vermont General Assembly, An act relating to data brokers and consumer protection., enacted 22 May 2018

Electronic Frontier Foundation's Recommendations for Consumer Data Privacy Laws, Gennie Gebhart, 17 June 2019

Twelve Million Phones, One Dataset, Zero Privacy, Charlie Warzel and Stuart A. Thompson, New York Times, 19 Dec 2019

Facebook and Cambridge Analytica, Bruce Schneier, 29 March 2018