March 16, 2016   |   Jake Anderson
March 16, 2016
(ANTIMEDIA) Imagine you’re at your local coffee shop, browsing the Internet. You decide to login, and after doing so, you realize you have a new friend request. You don’t know the person, but the face seems familiar. Casually glancing around the coffee shop, you suddenly see the face of the user.
You have just experienced Facebook’s forthcoming algorithm, which will attempt to use patented data mining tools to use location-based resources to connect users who frequently visit the same location.
“The plurality of factors can include at least one of an inferred locational proximity between the first user and the second user, a frequency of inferred meetings between the first user and the second user, a duration of each of the inferred meetings between the first user and the second user, or a pattern of occurrences of inferred meetings between the first user and the second user.”
In other words, Facebook will track IP addresses and device signatures on public Wi-Fi networks in order to determine how often two different people are in the same locality and how much time they spend there.
According to the patent, Facebook’s algorithm will employ broadcast triggers that include gyroscopes, accelerometers, and motion processors to track a variety of movements occurring on the Wi-Fi network. These movements include stationary patterns, walking, running, and vehicle-riding.
Facebook describes another example of the algorithm in action:
“The first user and a third user take the bus together every day over a certain time period and become acquainted. In this example, the first user later decides to begin driving instead of taking the bus. The first user wishes to connect with the third user but the first user does not have the third user’s contact information. The first user also does not have any other practical way of finding the third user within the social networking service. In this example, the first user and the third user are unable to find each other and reestablish communications. As such, it can be advantageous to provide an approach for users, who have met or have likely met, to connect with one another if they so choose.”
Going further, Facebook explains that ‘friends,’ (which, oddly, they say can refer to an “edge formed between and directly connecting two user nodes”) can make connections bilaterally or unilaterally, or “one-way.”
The meaning of this is still unclear, and Anti-Media was unable to obtain a statement from Facebook. We also reached out to the Electronic Frontier Foundation about whether the new app could entail a breach of privacy.
What A Future offers more analysis of the tech side of the algorithm:
“Whenever you’ll enter into a Wi-Fi network, Facebook’s app will launch a data packet (either by Bluetooth or NFC) containing a little info about you and your device. Your smartphone will also receive data packets released by other Smartphones on the same network. These packets play central role in tailoring the suggestions.”
The new algorithm uses what in Big Analytics-speak is called a federated database, which maps decentralized metadata into a single database management system. This data centric reference architecture is also used by LinkedIn and Twitter.
If you’re starting to feel creeped out and confused, What A Future says you shouldn’t be worried that everyone in the coffee shop will know who you are, claiming Facebook’s filtering mechanisms will work to make logical friend recommendations. The weird guy on the bus who’s always staring at you — just don’t confirm his friendship request. End of story, right?
Are you comfortable with the idea of a social network tracking your location in order to connect you with another user in your physical proximity? You’re about to find out.
This article (Facebook’s Next Algorithm Will Help You ‘Break the Ice’ with Strangers in Real Life) is free and open source. You have permission to republish this article under a Creative Commons license with attribution to Jake Anderson and theAntiMedia.org. Anti-Media Radio airs weeknights at 11pm Eastern/8pm Pacific. If you spot a typo, email email@example.com.