Facebook, the most-used social network in the world, is also one of the least understood.

A data juggernaut, Facebook has spent the better part of the last decade sucking up as much information about its users as possible — what they’ve liked, what websites they’ve visited, and even where they’ve shopped. What’s less clear is how Facebook uses that data to actually make decisions about which ads and other content it wants to show users.

What Facebook Thinks You Like, a new project from ProPublica, is trying to decode all of that. The tool, an extension for Google’s Chrome browser, let users see exactly what activities, brands and products Facebook, based on its data, thinks they like. The tool also tells users which — and how many — advertising categories those interests place them in.

This is journalism as Chrome extension and it’s great. https://t.co/sOZmmMY7IH

— Jake Grovum (@jgrovum) September 28, 2016

While Facebook already freely offers this data to users, ProPublica says its goals are not only to figure out how Facebook uses user data in its algorithm, but also to get a better sense of how people actually feel about all of it. To that end, it’s working with WNYC’s Note to Self, which is asking readers to send in their reactions to using the tool.

The tool is the first release in a new four-part ProPublica series called Breaking the Black Box, which will take a deeper look at the inner workings and biases of the secret algorithms used by the big tech companies. The organization has already spent the better part of the past year reporting on the topic, most recently with a deep dive on how Amazon uses its pricing algorithm to give it an edge over the competition.

The results are sure to be eye-opening for the many Facebook users still unaware of the sophistication and scope of Facebook’s data collection and how it’s used to target ads. On the other hand, for lots of early users, ProPublica’s tool has shown that while Facebook collects plenty of data, its resulting conclusions are often comically off target (or comically specific).

#Facebook thinks I like Dating, Soft Drinks and the Communist Party of China; couldn’t be more wrong. @ProPublica https://t.co/QOYvkdz448

— Myrthe Prins (@PrinsMyrthe) September 28, 2016

According to ProPublica’s Facebook tool, FB thinks I like “Bag” because “you clicked an ad related to Bag” pic.twitter.com/JWboC3MXfi

— Mathew Ingram (@mathewi) September 28, 2016

interesting take, facebook (check out @ProPublica‘s cool tool on FB’s algorithm here: https://t.co/mKfIuat4Ae) pic.twitter.com/7abPDv0U3h

— Andrew Heiss (@andrewheiss) September 28, 2016

ProPublica investigative reporter Julia Angwin appeared on WNYC’s Note to Self earlier today to talk about the project and ProPublica’s investigations.