Mangaki

Within a set of search results, what should you do to find the optimal solution for not just a single user but a group? Mangaki is building an open source library for privacy-preserving group recommendations of items. While many content providers suggest recommendations at a personal level, these are often directed to a single user, or are restricted to a generic “family” category. Whenever say a group of friends want to watch a movie, it is often hard to decide what to watch, because people can have really different tastes.

Recommendations are also very privacy-sensitive. A straightforward way might be to share our complete viewing history, but that certainly can lead to embarrassing and awkward situations. So how can we collectively compute a list of relevant items without disclose all of our data unencrypted. The Mangaki project is making an open source library for group recommendations that works in a scalable and distributed way.