In this proposal, we aim to develop a tool for end users to explore how their data is being aggregated by Facebook and its partners. In other words, our goal is to allow users to see what data Facebook knows about them and is making available as a targeting parameter to its advertisers, and whether that data comes from Facebook itself or a data broker. Our proposal has three primary tasks:
1) Measurement study of data brokers. We first plan on collecting a list of all of the Facebook partner data brokers, and the categories of data that they provide, by crawling Facebook’s ad submission web pages in countries around the world. Our preliminary data crawl shows that there are at least five different data brokers who partner with Facebook who collectively provide 35 categories of data (covering 1,111 different potential attributes), and that the data they provide varies significantly by country.
2) End-user tool to reveal Facebook and data broker information. We propose to then build a tool that will (a) show Facebook users the attributes that Facebook has on them that advertisers are able to target, and reveal the data partner from which this data came.
3) User survey for data accuracy. Finally, we propose to implement a user survey as part of the end-user tool. This survey will allow users to report back to us (in an anonymized fashion) about the completeness and accuracy of the data that Facebook and its partners have collected. The results of this survey will allow us to understand how much data Facebook is collecting worldwide.